sign in sign up
<<
Home , Adaptive mutation, Bacteria, Genome evolution, John Roth (geneticist)

To my family

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Pettersson ME, Sun S, Andersson DI, Berg OG. (2009). of new functions: simulation and analysis of the amplification model. Genetica 135(3): 309-324.

II Sun S, Berg OG, Roth JR, Andersson DI. (2009). Contribution of gene amplification to evolution of increased resistance in typhimurium. 182(4): 1183-1195.

III Koskiniemi S, Sun S, Berg OG, Andersson DI. (2012). Selection- driven reduction in . (Submitted)

IV Sun S, Ke RQ, Hughes D, Nilsson M, Andersson DI. (2012). High frequencies of genome rearrangements in bacterial . (Manuscript).

V Sun S, Negrea A, Rhen M, Andersson DI. (2009). Genetic analysis of colistin resistance in serovar Typhimurium. Antimicrob Agents Chemother 53(6): 2298-2305.

VI Sun S, Zhang W, Mannervik B, Andersson DI. (2012). Evolution of increased ß-lactam resistance in an engineered Metallo-ß-lactamase. (Manuscript).

Reprints were made with permission from the respective publishers.

Contents

Introduction ...... 9 Salmonella enterica ...... 9 and amplification (GDA) ...... 10 Dynamics of gene amplification ...... 10 Mechanisms of formation and loss of GDA ...... 11 GDA and adaptive evolution ...... 13 ...... 19 Mechanisms of deletion formation ...... 19 Reductive evolution ...... 21 Inversion ...... 22 Formation of inversions ...... 23 Constraints on inversions ...... 24 Antibiotic resistance ...... 24 Resistance mechanisms ...... 25 rates ...... 26 Spread of antibiotic resistance through HGT ...... 27 Fitness cost and compensatory ...... 28 Present investigations ...... 30 Gene amplification promotes innovation of new ...... 30 Gene amplification facilitates antibiotic resistance ...... 31 Adaptive genome reduction ...... 33 Detecting spontaneous genome rearrangements ...... 35 Genetic analysis of colistin resistance ...... 37 Evolution of a novel Metallo-ß-lactamase ...... 38 Connecting binding and β-lactamases ...... 39 Concluding remarks ...... 40 Future perspectives ...... 43 Chinese Abstract (摘要) ...... 45 Acknowledgements ...... 46 References ...... 48

Abbreviations

S. typhimurium Salmonella enterica serovar Typhimurium LT2 E. coli coli P. mirabilis mirabilis DNA Deoxyribonucleic acid RNA Ribonucleic acid mRNA Messenger RNA tRNA Transfer RNA rRNA Ribosomal RNA LPS ATP Adenosine triphosphate MRSA Methicillin-resistant aureus NDM-1 New Delhi metallo-β-lactamase 1 PBP Penicillin binding ESC Extended spectrum ESBL Extended spectrum β-lactamase MBL Metallo-β-lactamse GTA Gene transfer agents bp kb Kilo base pairs Mb Mega base pairs GDA Gene duplication and amplification IAD Innovation-amplification-divergence SGR Spontaneous genome rearrangements MMR Methyl-directed mismatch repair BER Base excision repair NER excision repair HGT MIC Minimum inhibitory concentration PCR chain reaction RCR Rolling circle replication PFGE Pulse gel electrophoresis LB Luria Bertani broth LA Luria Bertani

Introduction

In 1859 Charles Darwin for the first proposed the theory of evolution by in his “On the origin of ” (DARWIN 1859) and this revolutionary scientific idea has ever since formed the foundation for understanding the mechanisms of evolution. As one of the major driving forces of evolution, natural selection occurs when a population of individuals differs in their fitness, in other words, their ability to survive and reproduce, and the fittest individuals will become more common in the population. Random is another process that drives evolution, whereby chance events fix individuals in the population independent of their fitness (KIMURA 1983). This thesis is focused on the studies of adaptive evolution. is the evolutionary process whereby an becomes more fit in a particular environment. Although adaptive evolution often leads to complex genetic traits, the underlying process only requires two players: mutation and selection. Mutation is the ultimate source of genetic variation that natural selection works on; therefore determining the properties of mutations is fundamental to understanding the mechanisms of adaptive evolution. Featured by fast generation , large population sizes, and the ability to be stored in a frozen non-evolving state, have been used with great success in laboratory evolution experiments to unravel the genetic basis of adaptation (CONRAD et al. 2011; PORTNOY et al. 2011). In , microorganisms are unique for their ability to adapt rapidly to different environments. One significant example is the development of antibiotic resistance. In this thesis, I have investigated the mechanisms of bacterial adaptation to new environments using adaptive laboratory evolution. Different types of mutations were under investigation with a particular interest in genome rearrangements. The adaptation process was focused on the development of bacterial resistance to .

Salmonella enterica As one of the main workhorses of bacterial genetics, Salmonella enterica serovar Typhimurium strain LT2, hereafter referred to as S. typhimurium, was used as a model organism throughout this work. S. typhimurium is a rod- shaped gram-negative enterobacterium that in humans causes with symptoms such as and (HOHMANN 2001). In mice, S.

9 typhimurium causes a systemic that is similar to (MASTROENI and SHEPPARD 2004). All strains used in this work derived from the LT2 strain, which is less virulent due to a defective rpoS gene (SWORDS et al. 1997). With a fully sequenced genome (MCCLELLAND et al. 2001), the S. typhimurium LT2 strain has been extensively studied in and a wide range of tools are available for genetic manipulation. Together with the common features shared by microorganisms such as fast growth and large population size, this makes S. typhimurium one of the most convenient model for studies.

Gene duplication and amplification (GDA) Gene amplification has been observed in all three kingdoms of (DEVONSHIRE and FIELD 1991; DUNHAM et al. 2002; ROMERO and PALACIOS 1997; WONG et al. 2007) and is important both from a fundamental evolutionary perspective as a major source of gene novelty (BERGTHORSSON et al. 2007; OHNO 1970), and in as a significant contributor to phenotypic variability among individuals and many human (ALBERTSON 2006; BECKMANN et al. 2007; CONRAD and ANTONARAKIS 2007). Differing from other types of mutations, GDA is more similar to a regulatory response considering its high prevalence and intrinsic instability. With these two properties, GDA is of significant biological importance in adaptive evolution where it can (i) provide a direct solution to a selective problem, (ii) facilitate further adaptation, and (iii) create new gene functions.

Dynamics of gene amplification Much of our knowledge about gene amplification comes from the pioneering studies in (E. coli), S. typhimurium and (P. mirabilis) (ANDERSON and ROTH 1981; ANDERSON and ROTH 1977; HASHIMOTO and ROWND 1975; HILL et al. 1990; PERLMAN and STICKGOLD 1977; PETERSON and ROWND 1985; PETES and HILL 1988; ROTH et al. 1996). These studies not only demonstrated that GDAs are highly prevalent in eubacteria and can affect any gene in bacterial , but also showed that GDAs are intrinsically unstable and could disappear in just a few generations of growth when the selection pressure is relaxed. In an unselected bacterial population, the frequency of cells with a duplication of any specific chromosomal region varies between 10-2 and 10-5 depending on the particular genomic region and the most common ones occur between directly repeated ribosomal RNA (rRNA) via (ANDERSON and ROTH 1981). The sizes of duplicated regions can range from a few kilo base pairs (kb) to several mega

10 base pairs (Mb) (ANDERSON and ROTH 1981; KUGELBERG et al. 2006; NILSSON et al. 2006; SONTI and ROTH 1989; STRAUS and D'ARI STRAUS 1976). Based on the frequencies and approximate sizes of the amplified regions, it has been estimated that at least 10% of cells in an unselected bacterial culture contain a duplication somewhere in their (ROTH et al. 1996). assays have been used to measure the steady-state frequencies of duplications in S. typhimurium (ANDERSON and ROTH 1981; SONTI and ROTH 1989). However, this method may not be applicable for determining the rates of duplication formation. The frequency of duplications rapidly approach steady state in the population (REAMS et al. 2010) and therefore small populations of cells must be used in order to measure the initial formation rate, which makes it technically difficult to apply the transduction assay. Instead one can estimate the formation rates of initial duplications by using the steady-state frequencies and loss rates of duplications according to the formula kduplication = steady-state frequency × kloss. The segregation rates of duplications have been estimated to range from 0.013 to 0.156 per per generation (Paper I). Together with the determined steady-state frequencies, this suggests that the rate of formation of spontaneous duplications in S. typhimurium varies between 10-2 and 10-5 per cell per generation, whereas the base substitution rate was estimated to -11 -10 be 10 ~10 per cell per generation (HUDSON et al. 2002). Therefore, GDA is much more likely to be the initial response under selective pressures. In the absence of selection pressure, GDAs may confer fitness costs on the bacterial cells carrying them. To examine the potential fitness costs of GDAs, the relative growth rates of five genetic duplications constructed in S. typhimurium were measured in rich medium and were found to range from being indistinguishable from wild type to reduced by 20%. There was no correlation between the size of the duplicated region and the fitness cost indicating that the genetic content of the duplicated region is the main determining factor rather than the metabolic cost of making extra DNA. One possible explanation is the gene dosage imbalance hypothesis, which suggests that stoichiometric imbalances in macromolecular complexes confer the fitness cost (LIANG et al. 2008; PAPP et al. 2003; VEITIA 2004). In summary, GDAs are very frequent in unselected bacterial populations and this high prevalence renders GDAs likely to be the initial response under any selection pressure. On the other hand, the intrinsic instability and potential fitness cost will cause GDAs to disappear rapidly in the population after the selection pressure is relaxed.

Mechanisms of formation and loss of GDA With regard to the formation of initial duplications, two mechanisms have been observed depending on whether RecA is involved or not. In S.

11 typhimurium, the most frequently observed spontaneous tandem duplications are those having their endpoints within directly repeated rRNA operons, which suggests that the formation of such duplications occur by non-equal homologous recombination between these long, directly oriented repeats (ANDERSON and ROTH 1981; LEHNER and HILL 1980). This process would duplicate the region between the two direct repeats and lead to a novel junction sequence (Fig. 1). In addition to the rRNA operons, other naturally occurring long repetitive sequences, such as sequences, transposable elements and repetitive extragenic palindromic sequences, have also been observed at the join points of duplications (HAACK and ROTH 1995; JESSOP and CLUGSTON 1985; LIN et al. 1984; SHYAMALA et al. 1990). In a recent study, it was found that the steady state frequency of duplications in an unselected bacterial population was about ten-fold lower in a recA mutant (REAMS et al. 2010). This result indicates that (i) the formation of some duplications are RecA-dependent and are probably due to the un-equal exchange between long repeats on two sister chromatids, and (ii) a significant proportion of spontaneous duplications are formed in a RecA- independent manner.

Figure 1. Formation of a tandem duplication by non-equal homologous recombination between two directly oriented repetitive sequences R or R’ present on sister chromatids.

Additional evidence that RecA is not required for the formation of some duplications comes from the analysis of the homologies at the duplication junction points. RecA-involved homologous recombination requires at least 20-40 sequence to form a recombinogenic complex, which sets the minimum length of junction homologies for RecA-dependent homologous recombination. However, very short homologies (less than 20 bps) have been observed at the junction sequences of genetic duplications (ALBERTINI et al. 1982; EDLUND and NORMARK 1981; REAMS and NEIDLE 2004; TLSTY et al. 1984; WHORISKEY et al. 1987) and recombination between such short repetitive sequences is often RecA-independent (JONES et al. 1982; LOVETT et al. 1993; LOVETT et al. 1994). The possible RecA- independent processes include strand slippage during DNA replication or repair (TRINH and SINDEN 1993), pairing of single stranded regions at the replication fork (LOVETT et al. 1993) and ligation of DNA ends via DNA gyrase (IKEDA et al. 2004).

12 Once a duplication is formed, the duplicated region can be further amplified to high level via RecA-dependent recombination between the long perfect repetitive sequences that are generated by the initial duplication (REAMS and NEIDLE 2004). In addition to non-equal cross over, high-level amplification can also occur independent of the formation of duplication by mechanisms based on rolling circle replication (RCR) (PETIT et al. 1992), which could generate a large tandem array in a single generation (ANDERSSON et al. 1998a; GALITSKI and ROTH 1997). As aforementioned, GDA is genetically unstable and could be rapidly lost in the absence of selection on the amplified array. A number of studies have shown that GDA could be stabilized in recA mutants, which suggests that loss of GDA is mediated by RecA-dependent homologous recombination between different units of the amplified array (ANDERSON and ROTH 1978; GALITSKI and ROTH 1997; HAACK and ROTH 1995; HILL et al. 1969; LIN et al. 1984). However, deletion between repeated sequences also occurs in a RecA-independent manner in some system (BI and LIU 1994; LOVETT et al. 1993; MATFIELD et al. 1985).

GDA and adaptive evolution Bacterial adaptation to new environmental conditions mainly occurs via two different mechanisms: alteration of gene regulation without any inheritable genetic change or selection of novel adaptive phenotypes conferred by stable mutations. As previously discussed, GDA is probably the most frequent mutation type in bacterial populations, but meanwhile, GDA is intrinsically unstable. These two unique properties make GDA more similar to a regulatory response (Fig. 2). GDA plays several important roles in bacterial adaptation processes: (i) providing a direct solution by increasing gene dosage when bacteria face new challenges, (ii) facilitating further stable genetic adaptation, and (iii) establishing new gene functions. In addition to the three roles that GDA plays in adaptive evolution, I will also discuss the functions of GDA in a special evolutionary process, the development of antibiotic resistance.

13

Figure 2. Frequency in bacterial populations and response time for regulatory responses, GDAs and point mutations. Adapted from Andersson et al. (ANDERSSON and HUGHES 2009).

GDA provides a direct solution One selective problem that bacteria might frequently encounter under natural conditions is limited nutrient sources. GDA has been observed to arise frequently when bacteria are subjected to selections for growth on limiting sources and the growth advantage conferred by GDA is likely due to the increased dosage of genes involved in utilization of the limiting carbon sources (MCBETH and SHAPIRO 1984; NEUBERGER and HARTLEY 1981; OGAWA and MIYASHITA 1995; RANGNEKAR 1988; REAMS and NEIDLE 2003; SONTI and ROTH 1989). GDA has also been shown to confer antibiotic resistance in bacteria through over-production of antibiotic-modifying , target molecules and efflux pumps (BROCHET et al. 2008; CLEWELL et al. 1975; EDLUND et al. 1979; HAMMOND et al. 2005; MATTHEWS and STEWART 1988; NICOLOFF et al. 2006). One of the earliest examples demonstrating that high level gene amplification can be selected under strong selective pressure was the study of plasmid-borne ampC amplification by Normark and coworkers (EDLUND et al. 1979; EDLUND and NORMARK 1981). An E. coli strain with the ampC gene located on a plasmid was selected for elevated resistance to ampicillin. After several rounds of selection, a 10 kb region including the ampC gene was found to be amplified up to 50 fold leading to a 40-fold increase in resistance. Apart from allowing organisms to cope with external selection pressures as discussed above, GDA has also been demonstrated as a possible adaptive response to internal selection pressures, such as deleterious mutations. Antibiotic resistance mutations are commonly associated with a fitness cost in the absence of

14 antibiotic selection when compared with wild type (ANDERSSON 2003; ANDERSSON 2006; ANDERSSON and LEVIN 1999). When the antibiotic selection is removed, the resistant bacteria can regain the fitness by secondary compensatory mutations, which will often restore the fitness without loss of resistance. In one case, amplification of genes unrelated to the original resistance mutations was found to reduce the fitness cost of antibiotic resistance (NILSSON et al. 2006). The antibiotic used for selection in this work was actinonin, which inhibits by preventing removal of the formyl group from formyl-, the first incorporated in proteins. The mutations that disable the formylation functions substantially reduced bacterial fitness in the absence of antibiotic. One common type of compensatory mutation identified was amplification of the metZW genes, which restored the fitness by overproduction of nonformylated methionyl tRNAi. This example demonstrated that gene amplification might occur as a common adaptive response to various types of deleterious mutations.

GDA facilitates further stable genetic adaptation With its high prevalence, GDA can easily provide a direct solution to a selective problem; on the other hand, due to its intrinsic instability, probably in most circumstances, GDA functions as a facilitator to allow and population expansion and thus increase the probability of acquiring rare stable mutations. Finally, after the selection pressure is relaxed by formation of stable adaptive mutations, the parental amplification will segregate back to a single copy without leaving any trace in the genome. This process has been demonstrated in a number of studies with different experimental setups (ANDERSSON et al. 1998a; HASTINGS et al. 2000; PAULANDER et al. 2010; PRANTING and ANDERSSON 2011; ROTH et al. 2006). One classical and extensively studied case is the amplification model proposed to explain an interesting but controversial phenomenon, lac adaptive mutation in the Cairns system. The experimental system was originally designed to study the mutations that are stress-induced (CAIRNS et al. 1988). In the modified – system (CAIRNS and FOSTER 1991), Lac cells with a reversible and leaky +1 in the lac are plated on medium, which barely permits cell growth. About 100 revertants (Lac+) appear on a lawn of 108 non-growing parent cells over 6 days. Given that the mutation rate is less -8 than 10 /cell/division under non-selective conditions (ROSCHE and FOSTER 2000), it seems that the mutation rate is increased about 100-fold by the selection. Although several models have been proposed to explain this “adaptive mutation” phenomenon by stress-induced , the amplification model best satisfies the basic observations where selection- driven gene amplification of the leaky mutant lac operon facilitates acquiring a Lac+ reversion event (-1 frameshift mutation) without enhanced mutagenesis (ROTH et al. 2006) (Fig. 3).

15

Figure 3. Amplification model. The pre-existing duplications allow initial growth on lactose that is improved by further lac amplification. The high copy number of the lac operon rather than mutagenesis increases the probability of acquiring a reversion mutation.

In 108 plated cells, approximately 106 cells (1%) have a duplication or amplification of the lac operon that initiates slow growth on lactose medium (SLECHTA et al. 2003; SLECHTA et al. 2002). Within these many clones, the cells with increased mutant lac operon copy number will be progressively selected due to their growth advantage. Thus, both the cell growth and increased copy number of the lac operon enhance the probability of a reversion event (-1 frameshift). When a reversion event occurs, one functional lac+ allele will relax the selection pressure on the amplified array and the subsequent segregation will generate the haploid Lac+ cells. Finally, these haploid revertant types overgrow the original colony. In summary, the amplification model described above and the related work demonstrate how GDA facilitates the cells to escape from growth limitations without increasing the rate of mutagenesis.

GDA creates novel gene functions Gene duplications are generally thought to be the main source of new genes and fundamental to increasing organism complexity. The original ideas on creation of new genes were developed by Ohno: duplication creates a

16 redundant gene copy that is free from selective constraints and acquires a new function by accumulating beneficial mutations while the other retains the original function (OHNO 1970). A problem with this classic model is that due to its supposed neutrality the newly duplicated gene is subject to loss by inactivating mutations, drift, and gene conversion and the only way for the duplicated gene to be permanently preserved is by the of rare beneficial mutations that elicit a new beneficial function. In addition, the assumption that duplications are neutral might not always hold due to the possible metabolic cost and gene dosage effect (PAPP et al. 2003; REAMS et al. 2010; VEITIA 2002; VEITIA 2004). In bacteria the main problems are likely to be segregational loss and counter-selection given the observed high loss rate (up to 10% per generation) and varying fitness cost (from undetectable to 15%) of duplications (Paper I). Thus, in this model loss is expected to be the most common fate of duplicated genes (LYNCH and CONERY 2000; LYNCH and FORCE 2000). However, these theoretical expectations are not fully consistent with the observed high frequency of duplicate-gene preservation (20%-50%) following genome duplication events for long periods of time (AMORES et al. 1998; FERRIS and WHITT 1979; NADEAU and SANKOFF 1997; WENDEL 2000). This suggests that the duplicated genes must be subject to some type of selective constraint that purges null alleles and preserves the duplicated gene. Several models have been proposed to explain the preservation of duplicate genes: (i) redundant genes might be favored by selection because it protects genomes from degenerative mutations that have negative effect on fitness (CLARK 1994; NOWAK et al. 1997); (ii) gene duplicates are preserved by sub- functionalization, whereby both copies are stabilized by mutations that inactivate one sub-function of each copy (FORCE et al. 1999; LYNCH and FORCE 2000); (iii) if heterozygotes at a single locus confer a higher fitness than homozygotes (over-dominance), this advantageous can be stabilized by duplication of the locus followed by recombination that combines the two alleles in the same haploid genome (HALDANE 1933; OHNO 1970). Although the second and third models described above can explain how the duplicated genes might be preserved, they do not explain how a totally new function is acquired because both copies are under selection to provide the original function or fitness advantage. The innovation-amplification-divergence (IAD) model was proposed recently to explain how GDA creates new gene functions (BERGTHORSSON et al. 2007) (Fig. 4). In the first step (innovation), the original gene does not only have its primary function (designated as function A in Fig. 4), but also secondary minor functions (designated as function B in Fig. 4) that are neutral at the beginning. The new gene formation process is initiated when the secondary minor activity is favored by selection in a new environment. In the second step (amplification), the parental gene is duplicated and further amplified to high copy number due to selection of an increase in the secondary function

17 B. In the third step (divergence), the extra gene copies enhance the probability of acquiring a mutation that improves the secondary function because there are more targets for mutations. When selection on the amplified array is relaxed by the improved B function, the amplification will segregate leaving one copy with the new function B and one copy with the original function A. There are two prerequisites for this model to work efficiently: (i) GDA is common. This prerequisite is fulfilled very because gene duplications are among the most common types of mutations in bacteria, as was discussed in the previous section “Dynamics of gene amplification”; (ii) secondary minor activities exist in the protein encoded by the parental gene. Recently, systematic protein overexpression studies have demonstrated that promiscuous activities are common in proteins and gene overexpression can make such activities available for selection (ANDERSSON 2011; PATRICK et al. 2007; SOO et al. 2011).

Ancestral protein with secondary activity

1) Innovation AB B-function is favored by selection in new environment

Segregation Duplication

AB AB

Duplication and amplification 2) Amplification Segregation Amplification are stabilized in population by selection for B-activity

(AB)n = “B”

AB AB n

Mutation leading to better B-function

AB AB AB 3) Divergence n-1 Improved B-function relaxes 3) Divergence Segregation the selection pressure and leads to segregation of the amplified arrays AB AB

Figure 4. The innovation-amplification-divergence (IAD) model.

18 Deletion As opposed to gene duplication and horizontal gene transfer (HGT) that increase DNA content, deletion leads to loss of DNA content and is an irreversible genetic event. The small and compact bacterial genomes of obligate intracellular parasites and are thought to be due to deletional bias (where deletions are more frequent than insertions) counteracting genome expansion led by frequent HGT events (MIRA et al. 2001). This process can occur via either large deletions that can remove multiple genes in a single event or small deletions that erode generated by inactivating mutations. Two central requirements for deletional bias to lead to gene loss are thought to be relaxed selection and increased genetic drift (ANDERSSON and KURLAND 1998; MORAN 2002; MORAN 2003; MORAN and 2004; MOYA et al. 2008). However, these findings do not exclude the possibility that genome reduction is an adaptive process, whereby deletions are beneficial for bacteria and favored by selection.

Mechanisms of deletion formation Tremendous efforts have been made in elucidating mechanisms of deletion formation, which have generated different intriguing explanatory models and greatly improved our understanding of the genetic basis of deletion formation. Based on the existence and length of homologous sequences at the recombination sites, deletions can be formed via either illegitimate recombination with or without short homologies, or via homologous recombination. The genetic properties of deletion formation between long homologies have been extensively studied by using artificially constructed tandem repeats (~600-800bp) on or chromosomes (BIERNE et al. 1997b; BIERNE et al. 1997c; BZYMEK and LOVETT 2001; LOVETT et al. 1993; SAVESON and LOVETT 1997). Although the majority of these deletion events require a functional RecA protein and presumably occur via RecA- dependent homologous recombination pathways (KOWALCZYKOWSKI 2000), a substantial fraction was found to be RecA-independent. Two replication- based models were proposed to explain the RecA-independent deletion of repetitive sequences: (i) the simple slipped misalignment model and (ii) the sister-chromosome exchange associated slippage model (BZYMEK and LOVETT 2001). The first model was originally proposed to explain frameshift mutations (STREISINGER et al. 1966). In this model, the deletion is formed by misalignment between the newly synthesized strand and template strand in one sister chromosome leaving the other sister chromosome unchanged (Fig. 5A). Thus, deletions are not formed with concomitant expansions and therefore cannot account for those deletion events that are associated with dimerization. The second model was proposed to explain these dimeric products. In this model, the two nascent

19 strands of the sister chromosomes are both involved in the misalignment process, wherein a Holliday junction is formed and resolved possibly leading to a crossover (Fig. 5B). The genetic basis and features of these two models have been reviewed in Bzymek and Lovett 2001.

Figure 5. (A) Replication slippage model. Adapted from Bierne et al. (BIERNE et al. 1997c). (B) Sister-strand exchange model. Adapted from Bzymek et al. (BZYMEK and LOVETT 2001). (C) Gyrase subunit exchange model. Adapted from Ikeda et al. (IKEDA et al. 1982). (D) Single-strand annealing model.

20 The deletion formation via short-homology independent or dependent illegitimate recombination is thought to be largely RecA-independent with some exceptions (ALBERTINI et al. 1982). The break-join model is favored in explaining illegitimate recombination with different versions depending on whether short homology is required or not. DNA topoisomerase has been shown to mediate strictly non-homologous processes in several studies (ASHIZAWA et al. 1999; BIERNE et al. 1997a; IKEDA et al. 1982; SHIMIZU et al. 1997; SHIMIZU et al. 1995). In this topoisomerase-mediated break-join model, the cutting and ligation reactions normally concomitantly performed by topoisomerase are uncoupled, which allows using an ectopic substrate leading to formation of recombinant DNA (Fig. 5C). For short-homology dependent processes, a double strand break between short homologous sequences is processed by exonucleases to generate two single-stranded ends, which allows short-homology promoted annealing and subsequent ligation (Fig. 5D) (BIERNE et al. 1997a; KONG and MASKER 1994; SHIMIZU et al. 1997; SHIRAISHI et al. 2005). Palindromic sequences, which are prone to form hairpin secondary structures, have been shown to facilitate deletion formation (EGNER and BERG 1981; GLICKMAN and RIPLEY 1984; GORDENIN et al. 1992; TRINH and SINDEN 1991; WESTON-HAFER and BERG 1989). In these cases, the hairpin structures potentially juxtapose deletion endpoints, which promotes slippage of template strand and misalignment with nascent strand during replication presumably facilitated by the short homologous sequences flanking the palindromic sequences.

Reductive evolution Bacterial genomes can vary in size by 80-fold (from 0.16 Mb for the smallest known genome to up to 13 Mb for large genomes) (CASJENS 1998; NAKABACHI et al. 2006; SCHNEIKER et al. 2007). It was once thought that larger genomes evolved from smaller genomes by repeated genome duplication events based on the distribution of genome sizes of different bacterial species and the relative positions of duplicated genes (RILEY and ANILIONIS 1978). However, these two arguments are weakened by later findings that (i) the variation within a bacteria species is often very large (BERGTHORSSON and OCHMAN 1995), (ii) related bacteria with similar genome sizes have very different gene contents showing that gene duplications occur subsequent to the divergence of bacterial species (HUYNEN and BORK 1998; JORDAN et al. 2001). Furthermore, phylogenetic analyses demonstrate that endosymbiotic bacteria with smaller genomes evolve from free living bacteria with larger genomes via reductive evolution (ANDERSSON and KURLAND 1998). The increased availability of fully sequenced genomes of obligate endosymbionts or intracellular has allowed extensive genome- based studies of the reductive evolutionary processes that occur during

21 adaptation (ANDERSSON et al. 1998b; COLE et al. 2001; FRASER et al. 1997; FRASER et al. 1995; SHIGENOBU et al. 2000). Based on the comparative genomic studies, reductive evolution is thought to be driven by the following factors: (i) increased levels of genetic drift due to small effective population sizes (ANDERSSON and KURLAND 1998; MCCUTCHEON and MORAN 2012; MORAN 1996; WERNEGREEN and MORAN 1999), (ii) relaxed selection in the host environment (MORAN 2003; MORAN and PLAGUE 2004; MOYA et al. 2008), (iii) an underlying mutational bias towards deletions (MIRA et al. 2001), (iv) and restricted horizontal gene transfer associated with an intracellular life style (MORAN and PLAGUE 2004; MOYA et al. 2008). The loss of DNA could occur by large deletions spanning multiple genes (KATO- MAEDA et al. 2001; OCHMAN and MORAN 2001) and by small deletions that gradually erode inactivated genes (ANDERSSON and ANDERSSON 1999; ANDERSSON and ANDERSSON 2001). Which one is the dominant route will vary between different bacterial species. Sequence comparison between and a reconstructed enteric indicated that large deletions including as many as 50 genes occurred during the early stages of genome reduction (OCHMAN and MORAN 2001). On the other hand, as much as 24% of the prowazekii genome consists of non-coding DNA and pseudogenes speculated to be remnants of inactivated ancestral genes, which suggest that reductive evolution occurred by gene inactivation and subsequent small deletions (ANDERSSON and ANDERSSON 2001). Although these findings suggest that the main driving force in reductive evolution is genetic drift, the alternative hypothesis that this evolutionary reductive process is driven by selection is conceivable and has rarely been subject to experimental examination.

Inversion Recombination between inverted repeats can invert the intervening sequence and lead to another type of rearrangement, an inversion. Some types of inversions, such as the inversions of two specific DNA segments controlling alternative expression of two different types of subunit proteins in S. typhimurium (KUTSUKAKE et al. 2006), can be used as genetic switches allowing alternative expression of different gene sets. However, the most common large inversions in bacteria have more profound effects on bacterial phenotypes in terms of changing chromosomal structures and gene orders. Unlike duplications that are genetically unstable, inversions are more likely to cause permanent changes on bacterial chromosomes given that inversions are largely irreversible. E. coli and S. typhimurium have diverged from their common ancestor about 140 million years ago (OCHMAN and WILSON 1987) and the order of orthologous genes are strongly conserved between these two genomes in spite of about 15% sequence divergence (KRAWIEC and RILEY

22 1990). In contrast to this conservation, some inversions, especially those between the rRNA operons, are commonly observed during growth in culture (ROTH et al. 1996). Therefore, there must be strong constraints that either physically restrict the formation of inversions or counter-select the formed inversions due to their functional consequences.

Formation of inversions Most pioneering studies of inversion formation were based on placing sequences in inverse order at known chromosomal positions and examining inversions formed by recombination between these inverted repeats (KONRAD 1977; MAHAN and ROTH 1988; REBOLLO et al. 1988; SEGALL et al. 1988; SEGALL and ROTH 1989). These induced inversions were thought to be formed by either a full exchange between inverted repeats in the same chromosome or two half exchanges between sister chromosomes (Fig. 6) and these recombination processes were largely dependent on the function of the RecA protein (SEGALL and ROTH 1994).

Figure 6. Formation of inversions between long repeats in opposite orientations. Adapted from Roth et al. (ROTH et al. 1996).

Genome comparison studies have identified large chromosomal inversions with endpoints within or proximal to oppositely-oriented large repetitive sequences, such as rRNA operons and insertion sequences (ALOKAM et al. 2002; DARLING et al. 2008; HILL and HARNISH 1981), which suggests that

23 RecA-dependent homologous recombination has played an important role in formation of these large inversions. However, it has been shown that inversions can also be mediated by inverted repeats in a RecA-independent manner and replication-based models were proposed to account for the observed recombination products (BI and LIU 1996; TILLIER and COLLINS 2000).

Constraints on inversions Although it was suggested that mechanistic limitations could prevent particular sites from being recombined to form inversions (MIESEL et al. 1994; SEGALL et al. 1988), numerous studies have demonstrated that large inversions are commonly deleterious to bacterial cells and the fitness costs are caused by the functional consequences of the inversions (CAMPO et al. 2004; ESNAULT et al. 2007; HILL and GRAY 1988; HILL and HARNISH 1981; LIU et al. 2006; REBOLLO et al. 1988). The possible functional consequences of detrimental inversions are: (i) inter-replichore inversions could disrupt the balance between two replication arms and interrupt coordinated bidirectional replication (ESNAULT et al. 2007); (ii) intrareplichore inversions change gene dosage because oriC-proximal genes are present in more copies during replication and have increased expression (LIU and SANDERSON 1996; SCHMID and ROTH 1987); (iii) intrareplichore inversions could change the transcriptional direction of those highly transcribed genes, such as rRNA genes, leading to collision between RNA polymerase and DNA polymerase (BREWER 1988).

Antibiotic resistance Evolution is often thought to occur gradually over a long time frame. However, this is not the case for the evolution of antibiotic resistance in bacteria, which has occurred surprisingly rapidly and poses a serious health threat. Soon after the discovery of the antibiotic penicillin in 1928, many other antibiotics with both natural and synthetic origins have been developed and commercially produced in large scale. Since then the overuse and misuse of these potent drugs have exerted a very strong selective pressure on bacterial populations. Bacteria are well known for their ability to adapt to any new or changing environment and we are now facing a long list of microbes that have found their ways to circumvent the effects of different classes of antibiotics. A notorious case is the methicillin-resistant (MRSA), which has evolved resistance to β-lactam antibiotics including and . Although most MRSA strains remain susceptible to , vancomycin-resistant MRSA has also emerged by acquisition of resistant determinants from vancomycin-

24 resistant enterococci via horizontal gene transfer (LEVY and MARSHALL 2004). The emergence of multi-drug resistant gram-negative bacteria, notably strains of E. coli, pneumoniae, aeruginosa and baumanii, have become an even more serious threat in hospital settings (LIVERMORE 2004). Recently, the New Delhi metallo-ß- lactamase 1 (NDM-1) producing gram-negative bacteria has attracted significant attention due to their resistance to carbapenem antibiotic, which is often considered to be the last line of defense against bacterial (KUMARASAMY et al. 2010). Due to the ever-increasing problem of antibiotic resistance and an essentially dry antibiotic development pipeline, more efforts are required to understand the underlying mechanisms of antibiotic resistance.

Resistance mechanisms Bacteria have acquired different resistance mechanisms in order to deal with the strong antibiotic pressure. Common resistance strategies include enzymatic modification of drugs, mutating drug targets, enhancing efflux pump expression, and decreasing membrane permeability (ALEKSHUN and LEVY 2007) (Fig. 7). For example, ß-lactam antibiotics, which inhibit biosynthesis in bacteria by irreversibly binding to the PBP active site, have been extensively used against bacterial infections since the discovery of penicillin. This has led to the selection and spread of many different types of resistance mechanisms (LIVERMORE 1995; WELDHAGEN 2004). Prominent among these mechanisms is a class of enzymes, ß-lactamases, that can hydrolyze the ß-lactam ring in the antibiotic (AMBLER 1980; POOLE 2004). As bacteria developed resistance to early generations of ß-lactams, stable extended spectrum cephalosporins (ESCs) were introduced, leading to evolution of extended spectrum ß-lactamases (ESBLs) in bacteria (PETROSINO et al. 1998). Other bacterial resistance mechanisms to ß-lactams involve reduced permeability due to reduced levels of certain outer membrane proteins, altered target proteins (PBPs) or acquisition of novel PBPs that have a reduced binding affinity for ß-lactams, and expression of efflux pumps that extrude the antibiotics (AMYES 2003; MARTÍNEZ- MARTÍNEZ 2008; POOLE 2004; ZAPUN et al. 2008).

25

Figure 7. Mechanisms of bacterial resistance to antibiotics.

Mutation rates Bacteria can develop antibiotic resistance by acquisition of resistance genes through horizontal gene transfer (HGT) or de novo chromosomal mutations. Spontaneous mutations arise from various sources, including replication errors, oxidatively damaged bases, and endogenous DNA lesions (e.g., deamination, depurination, and methylation) (MAKI 2002). In addition, exogenous lesions caused by the can also lead to mutations. However, spontaneous mutations occur at very low frequency in living cells. For example, the base substitution rate in E. coli is often on the -10 order of 10 or lower per base pair per cell per generation (FOWLER et al. 1974; MACKAY et al. 1994). Thus, bacteria must have evolved efficient repair systems to avoid the generation of spontaneous mutations. The four major types of repair mechanisms include proof reading function of DNA polymerase, methyl-directed mismatch repair (MMR), the nucleotide excision repair system (NER), and different pathways of base excision repair (BER) (FRIEDBERG et al. 1995; LINDAHL and WOOD 1999). The proof reading function is supplied by the exonuclease activity of DNA that edit out incorrectly paired bases (ECHOLS and GOODMAN 1991). MMR efficiently corrects replication errors by recognizing the hemi- methylated 5’-GATC-3’ sequences, excising the mismatched bases from the unmethylated nascent strand and restoring the correct bases (MODRICH 1991). NER and BER mainly eliminate spontaneous DNA lesions before replication. Mutators are mutant strains that have increased mutation rates due to disruption of DNA replication or repair functions (MILLER 1996). Their existence, together with antimutators that have reduced mutation rates, clearly demonstrate that genomic mutation rate is genetically adjustable and therefore can be influenced by natural selection (MILLER 1996; SCHAAPER

26 1998). In a seminal paper, Drake found a nearly constant rate of genomic mutations in diverse organisms (DRAKE 1991). So what evolutionary forces are responsible for the observed uniform genomic mutation rate? Newly arisen mutations can be deleterious, neutral or beneficial in terms of their impacts on the fitness of organisms. Since the vast majority of mutations are likely to be deleterious, there will be a continued selection for decreased mutation rate (IMHOF and SCHLOTTERER 2001; KIBOTA and LYNCH 1996). Two selective forces have been suggested to prevent mutation rate from evolving to zero: (i) the need for beneficial mutations that favors higher mutation rates, and (ii) physiochemical or physiological cost of reducing mutation rates in the organisms (DRAKE 1991; SNIEGOWSKI et al. 2000). In asexual populations, it is likely that these three forces work together to modulate the genomic mutation rate that best promotes adaptation. In any bacterial population, mutators with increased mutation rates are expected to exist in low frequency mostly owing to spontaneous mutations in mismatch repair genes. For example, in non-selected E. coli populations the frequency of mismatch repair-defective mutators was estimated to be less -5 than 3×10 (BOE et al. 2000; MAO et al. 1997). During long time evolution, mutator phenotypes are likely to be disfavored by selection due to their increased genetic load (DE VISSER 2002; DENAMUR and MATIC 2006). Although mutators occur spontaneously at low frequency and might also start off with a disadvantage compared to non-mutators, they have been found to be able to reach high frequencies and take over a population both in laboratory settings and in nature (COOPER and LENSKI 2000; GROSS and SIEGEL 1981; JYSSUM 1960; SNIEGOWSKI et al. 1997). In most cases mutators are selected because they more rapidly produce beneficial mutations than wild type cells with normal mutation rates (DENAMUR and MATIC 2006). In the case of antibiotic resistance, mutators are expected to be favored because the selection for resistance mutations is much stronger than the counter-selection for the genetic load. There are examples providing clear evidence that antibiotic treatment can enhance the selection of mutators (GIRAUD et al. 2002; MAO et al. 1997).

Spread of antibiotic resistance through HGT Bacteria can acquire DNA from the environment or their neighbors through a process called horizontal gene transfer (HGT) and this process has played an important role in bacterial adaptive evolution (WIEDENBECK and COHAN 2011). In the case of bacterial resistance to antibiotics, HGT promotes this rapid evolutionary process by facilitating the spread of resistance genes within and between bacterial species. In fact the early evidence of the importance of HGT in bacterial evolution came from the observation that penicillin resistance spread among the (DATTA and KONTOMICHALOU 1965). HGT is mediated by at least four mechanisms:

27 transformation, conjugation, transduction and gene transfer agents (POPA and DAGAN 2011). Transformation involves uptake and integration of DNA from the environment and a physiological state, known as “competence”, is required for this process (CHEN et al. 2005; THOMAS and NIELSEN 2005). Some bacterial species, such as subtilis, pneumoniae, gonorrhoeae, and influenzae, are naturally competent. Conjugation is a process mediated by cell-to-cell junctions and a tunnel through which DNA can pass (WOZNIAK and WALDOR 2010). Plasmids are usually involved in conjugative transfer systems. Transduction occurs after phage infection. The transfer of DNA can occur either by imprecise excision of the phage genome from bacterial chromosome in specialized transduction or erroneous packaging of host DNA instead of phage DNA in generalized transduction (BRUSSOW et al. 2004). The fourth HGT mechanism is via gene transfer agents (GTAs). GTAs are cryptic or defective that are incapable of self-replication and tend to package fragments. This process is similar to generalized transduction (LANG and BEATTY 2007; STANTON 2007). Several different DNA elements have been described transferring antibiotic resistance genes: plasmids, transposons, and integrons ( et al. 2005; NORMARK and NORMARK 2002). Plasmids can be transmitted from one cell to another by conjugation, transformation or transduction. Resistance genes often reside on transposons or integrons that are integrated in plasmids (BARTH et al. 1976; BENEDIK et al. 1977; HALL et al. 1991; LUPSKI 1987). Transposons are that encode enzymes that mediate intracellular transpositions. These transposition- catalyzing enzymes are similar to those described for the insertion of phage genomes into bacterial chromosomes (MIZUUCHI and BAKER 2002). An integron consists of an integrase that can assemble tandem arrays of gene fragments and a upstream of the insertion site allows the expression of integrated genes (FROST et al. 2005). Another unique type of mobile genetic element associated with antibiotic resistance is the conjugative transposon, which can form a plasmid-like circular transfer intermediate that is transferred by conjugation (SALYERS et al. 1995).

Fitness cost and compensatory mutations Antibiotic resistance typically causes perturbations in the activity of cellular processes that often have deleterious effect on cellular fitness (ANDERSSON 2006) (Fig.8). But there are examples where resistance mutations confer very small or even no cost and it has been suggested that high-fitness resistance mutants are preferentially selected in clinical settings (BJORKMAN et al. 1998; BOTTGER et al. 1998; ENNE et al. 2005; O'NEILL et al. 2006). The fitness costs of resistance mutations are usually quantified as reduced growth rates in vitro or in . Both experimental and theoretical data

28 suggest that the biological fitness cost of resistance is a major determinant of the rate of resistance development (ANDERSSON 2003; ANDERSSON and LEVIN 1999).

Figure 8. Fitness cost of resistance mutations and compensatory evolution. S, R and C stand for susceptible, resistant and compensated bacteria respectively.

When selection is removed, antibiotic resistant bacteria can regain fitness either by reversion of the resistance mutations, which will always cause loss of resistance, or by secondary compensatory mutations that can partly or fully restore the fitness, often while maintaining resistance (BJÖRKMAN et al. 2000; MAISNIER-PATIN and ANDERSSON 2004; MAISNIER-PATIN et al. 2002) (Fig. 8). The frequency of secondary compensatory mutations is usually higher than reversion owing to the typically larger target size for compensatory mutations. In small populations, not all genetic variants are represented and thus compensatory mutations will be favoured over reversion because of their higher numbers. However, if the population size is large enough, the restored level of fitness will become the dominant factor and reversion will be favoured by selection.

29 Present investigations

Gene amplification promotes innovation of new genes Gene duplications are the main sources of new gene functions in genomes. The original ideas on creation of new genes were developed by Ohno: duplication creates a redundant gene copy that is free of selective constraints and evolves towards a new function by accumulating beneficial mutations (OHNO 1970). However, how the newly duplicated genes are maintained in the population for a sufficient time is still an unresolved issue for this classical model. Based on previous work on gene duplication and amplification in bacteria, the amplification model was proposed to account for the evolution of new genes (BERGTHORSSON et al. 2007). In this model, the new gene function is suggested to arise from a weak promiscuous activity that is increased by gene duplication and amplification and subsequently improved by point mutations in the extra gene copies. In paper I, we examined the role of gene amplification in creation of new gene functions by formalizing it into a mathematical framework. The simulated model consists of a population of cells, each of which carries a given number of copies of the gene and each gene copy has two properties that represent the activities of the old and new function respectively. As the cells replicate, the copy number of the gene is changed by recombination and the activity of each gene copy is changed by point mutations. The rates of these events depend on the rate constants for recombination, mutation and the number of gene copies. Simulations have allowed us to examine the dynamic process of the establishment of a new gene function. In the initial state, all cells carry a single copy of the gene with only the old function. Over time spontaneous duplication will occur in some cells. In most cases these duplications will segregate to a single copy, but once the cell with a duplication acquire mutations that create some activity of the new function, the copy number of the gene is likely to be increased by homologous recombination due to the positive gene dosage effect. As the new function becomes optimized, copy numbers will again decrease because the maintaining selection is relaxed and as a result carriage of multiple copies of the new gene becomes disadvantageous. Eventually the new gene function is completely optimized and a single copy will suffice to provide the new function. At this stage, cells with one copy of the original gene and one copy of the new gene will dominate the population.

30 The results of the simulations recapitulated several important features of the amplification model. Firstly, the amplification model is efficient only when the cost of the duplication of one gene is lower than the increased fitness acquired by a single mutation. In other words, the relationship between the potential fitness gain and the load is a key feature of the system. Increasing the load penalty has a strong detrimental effect on the amplification mechanism. Secondly, high recombination rates slow down establishment of new gene functions. This is because a high recombination rate leads to a high loss rate of unselected duplications, which both reduces the fraction of duplications present in the population and increases the chance that a copy with a beneficial mutation is eliminated before it spreads. However, the speed of the core amplification mechanism is reduced by very low recombination rates and is therefore also sub-optimal to the system. Thirdly, if one gene carries or can easily acquire a weak secondary function, the amplification of this gene can be stabilized by selection, which will facilitate establishment of a new function from this gene. The stabilized amplification provides more copies that increase the chance of appearance of a beneficial mutation and also preserve the original function. The parameter values used in this study were estimated from experimental data. The recombination rate in S. typhimurium was estimated from the segregation rate of a set of genetically constructed duplications. The duplications were stabilized by antibiotic selection on the resistance gene located at the join point of the gene duplicates. The duplications started to segregate when the antibiotic was removed. The estimated values of krec vary between 0.013 and 0.16. The rates of the same duplication-carrying S. typhimurium strains were measured to determine their fitness. The results showed that there was no obvious correlation between the fitness cost and the size of the duplication, indicating that one of the main prerequisites of the amplification model, duplications with low cost, can be achieved. But this also indicated that many duplications are not suitable for amplification due to high fitness costs. The steady-state duplication frequencies were determined at six different chromosomal locations for wild type S. typhimurium grown in LB as described previously (ANDERSON and ROTH 1981; REAMS et al. 2010). Together with the determined segregation rates, this suggests that the rate of spontaneous duplications in S. typhimurium varies between 10-2 and 10-5.

Gene amplification facilitates antibiotic resistance The extensive use of ß-lactam antibiotics has led to the evolution and spread of many chromosomal-, plasmid-, and transposon-borne resistance mechanisms (LIVERMORE 1995; WELDHAGEN 2004). TEM-1 ß-lactamase, encoded by the blaTEM-1 gene, can hydrolyze both penicillin and early

31 cephalosporins (MATAGNE et al. 1990). It has been previously shown that E. coli can develop ampicillin resistance by amplifying the ampC gene (EDLUND and NORMARK 1981). Gene amplification has been observed in both eubacteria and (CRAVEN and NEIDLE 2007; WONG et al. 2007) as response to various selective pressures, including antibiotics (ANDERSSON and HUGHES 2009; SANDEGREN and ANDERSSON 2009). Considering the relative high rate of duplication formation (ranging from 10- 2 to 10-5/cell/division) (paper I) and low rate of base substitution (~10- 10 /cell/division) (HUDSON et al. 2002), an increased level of any enzyme activity is more likely to be caused by a gene copy number change than by a point mutation. In paper II, the importance of gene amplification in bacterial adaptation to antibiotic resistance was investigated by experimentally evolving bacterial populations under selection of cephalosporins. A hybrid operon was constructed to contain both the blaTEM-1 gene and the gfp gene expressed from a single constitutive promoter (Fig. 9). This operon was placed on a conjugative F’128 plasmid. Compared with the control strain, the tester strain showed a small but reproducibly higher MIC against two of the tested cephalosporins, and cephalothin, which were chosen as the selective agents in this study. The evolution experiments were performed by plating independent lineages of the test strain (only for the 1st passage) or the selected mutant strains on a succession of agar plates containing increasing levels of cephalosporins. After five passages, the resistance level increased rapidly. In an initial selection, all 51 clones isolated were examined by real-time PCR and E-test to determine blaTEM-1 copy number and MIC. 70% of the clones showed an increased blaTEM-1 copy number, demonstrating that gene amplification was the most common response to selection.

Figure 9. Constructed hybrid operon on Tn10dtet transposon. The tetA gene encoding a tetracycline efflux protein and the tetR gene encoding a regulatory protein are the tetracycline resistance determinants. The gfp gene encoding a green fluorescence protein and the blaTEM-1 gene encoding TEM-1 β-lactamase are expressed from a single constitutive promoter called rpsM.

32 Clones isolated from six lineages evolved with cephalothin and two lineages evolved with cefaclor were genetically characterized in detail. DNA and mRNA levels of the blaTEM-1 gene were measured by real-time PCR and MICs were determined using E-tests. Of the eight lineages examined, four showed a similar evolutionary trajectory in which the MIC increase was paralleled by an increase in blaTEM-1 copy number. The other four had a similar pattern where the initial MIC increase was accompanied by an increased blaTEM-1 copy number, but subsequently followed by either constant or reduced levels of DNA and RNA. This indicated that later resistance-conferring mutations relaxed the selection pressure for the amplification and allowed it to segregate. To identify these secondary mutations, first, the blaTEM-1 gene and the rpsM promoter region were sequenced and no mutations were found. Second, transductants with Tn10dcam insertions linked to the increased resistance were screened and by Arbitrary PCR chromosomal locations of the linked Tn10dcam were determined. Two missense mutations in the envZ gene and the cpxA gene, and one deletion mutation in the nmpC gene were identified for three mutants from three different lineages respectively. The expression levels of the mRNAs encoding three major porin proteins were determined for envZ and cpxA mutants. These results showed that expression of these porin proteins was dramatically reduced, indicating that the cephalosporin resistance was conferred by reduced permeability of the outer membrane to cephalosporins (MARTÍNEZ-MARTÍNEZ 2008; NIKAIDO 2003; OPPEZZO et al. 1991). The mathematical modeling demonstrated that cells with some duplication or low amplification in place could rapidly expand the array to any copy number that is required to provide the necessary level of resistance. The present findings suggest that gene amplification facilitates the process of acquiring stable adaptive mutations by allowing a primary expansion of the population to a level that can support occurrence of rare point mutations, which in this case occur outside the amplified region.

Adaptive genome reduction Differing from duplications, which are highly unstable, deletions can remove multiple genes and are genetically irreversible. Several lines of evidence show that smaller genomes, such as bacterial obligate endosymbionts and intracellular pathogens have evolved from free-living bacterial species with large genomes (ANDERSSON and KURLAND 1998; BERGTHORSSON and OCHMAN 1995; HUYNEN and BORK 1998; JORDAN et al. 2001). The main mechanism is thought to be the fixation of neutral or deleterious deletions by increased genetic drift (due to population bottlenecks), in combination with relaxed selection and lack of horizontal gene transfer due to the obligate

33 intracellular life style (MCCUTCHEON and MORAN 2012; MIRA et al. 2001; MORAN 2002; MORAN and PLAGUE 2004). In paper III, an alternative hypothesis, selection-driven genome reduction, was subject to experimental examination. In order to determine the deletion rates and isolate spontaneous deletion mutants, we constructed a Tn10 transposon derivative that carries the lacZYA operon with a moaA and a resistance gene inserted into the lacA gene (Fig. 10). The constructed transposon was transposed into random positions on the chromosome and eleven strains with the inserted transposons were chosen to measure the local deletion rates. By simultaneously selecting loss of the moaA gene that confers chlorate resistance and loss of the lacZY genes that gives white colonies on McConkey agar plates, spontaneous deletions were detected. After being normalized to the experimentally identified deletable region, deletions rates were determined to vary between 5×10-12 and 1.25×10-9 per cell per generation per deletable kb of DNA, corresponding to a 225-fold difference in deletion rates between different chromosomal positions.

Figure 10. Engineered Tn10 transposon used for measurements of deletion rates. The tetA gene encodes a tetracycline efflux protein responsible for tetracycline resistance and the tetR gene encodes a regulatory protein. The lacY gene encoding β- galactoside permease and the lacZ gene encoding β-galactosidase are required for lactose catabolism. The lacI gene encodes a repressor protein of the lac operon. The moaA gene that encodes an enzyme involved in molybdate biosynthesis and a chloramphenicol resistance gene (cat) are inserted into the lacA gene.

The relative fitness of isolated deletion mutants was determined by measuring the exponential growth rates in rich/poor medium and/or competition experiments. 13/55 deletions isolated from three different chromosomal positions increased the fitness of the cells under at least one of three tested conditions, which indicated that beneficial deletions are common. To further test this hypothesis, six independent lineages of wild type S. typhimurium were grown by repeated for 1000

34 generations in rich LB medium and six different deletions were found after 1000 generations growth. These six deletions were then reconstructed in wild type background and two of them increased fitness with 4.7% and 3.2%, respectively. One deletion removed 5.2 kb of DNA including five genes (the uvrC, uvrY, yecF, sdiA, yecC, yecS genes) and the other was a 54bp small deletion in the fliG gene. One possible explanation for the observed beneficial deletions is that loss of these genes results in reduced or mass expenditure on DNA, RNA and protein. However, this viewpoint is less likely given that the estimated proportion of total amino acids for the deleted genes was much smaller than increased fitness. Since the deletion in FliG is expected to cause a defect in flagellar function (KIHARA et al. 2000), it is more likely that the increased fitness results from a reduction in ATP consumption to generate proton motive force that drives the flagella. In conclusion, our results suggest that reductive evolution can be driven by selection.

Detecting spontaneous genome rearrangements Genome rearrangements such as duplications, deletions and inversions have profound effects on bacterial phenotypes and the evolution of bacterial genomes. For example, duplications are the main sources of new gene functions and multiple stepwise deletions can lead to genome reduction. Most studies of genome rearrangements in bacteria have relied on the comparisons of closely related genomes (DARLING et al. 2008; SUYAMA and BORK 2001; THOMSON et al. 2008). However, the spectra of spontaneous rearrangements in homogenous and clonal bacterial populations are still poorly characterized. In unselected bacterial populations, spontaneous genome rearrangements (SGRs) are usually present at very low frequencies and therefore traditional technologies such as pulse-field gel electrophoresis (PFGE) and microarray-based comparative hybridization cannot be directly used to detect SGRs. In paper IV, we employed 454 pyrosequencing technology and a ‘split mapping’ computational method to detect SGRs by identifying their unique junction sequences. S. typhimurium was grown in a chemostat for 240 generations and bacterial cultures were collected at three time points (generation 48, 144 and 240) and used to prepare genomic DNA for . Collecting samples at different time points allowed us to examine how fast SGRs reach steady state frequencies. 454 pyrosequencing generated in total ~1 million reads of ~300 bases and the average sequencing coverage was calculated to be 63-, 48-, and 23-fold for the three samples. A read sampled across a rearrangement junction will leave a ‘split mapping’ signature in the reference, with a prefix and suffix of the read mapped to different locations. The relative chromosomal orientation and location of the prefix and suffix

35 were used to classify a putative rearrangement (Fig. 11). Reads with a ‘split mapping’ signature were selected using custom data mining methods and subsequently subjected to the confirmatory screening (Materials and Methods in paper IV). To experimentally verify the junction sequences of the potential SGRs, a new probing technique was employed in this work. This detection approach was based on padlock probes, which are designed for circularization when perfectly matching target sequences (NILSSON et al. 1994). Results based on a pilot detection experiment indicated that rearrangements with frequencies as low as 0.001% were detectable using this technique. Combining padlock probes and/or PCR (only used as an auxiliary confirmation method), we were able to confirm 22 unique junction sequences with more than 10bp junction homology, leading to an estimation of 28 small deletions, 12 duplications, and 11 inversions in the dataset for generation 48. Normalized by the average sequencing coverage, the frequency of SGRs was calculated to be around 40%, 20% and 20% for small deletions, duplications and inversions respectively. The deduced frequency for spontaneous duplications was in good agreement with previous estimates (ANDERSON and ROTH 1981). To our knowledge, neither inversion or deletion frequencies in bacteria have been measured previously on a genome-wide scale, thus our results provide new insights into frequencies of these two rearrangement events in bacterial populations.

Figure 11. Illustration of split mapping and rearrangement classification.

36 Genetic analysis of colistin resistance Colistin is a multicomponent , which was discovered in 1949, but was replaced in 1970s by antibiotics considered less toxic (LI et al. 2005; LI et al. 2006). However, the lack of effective antibiotics against multidrug-resistant gram-negative bacteria led to a revival in use of this antibiotic. The interaction between cationic colistin and negatively charged lipopolysaccharide (LPS) leads to a disturbance of the outer membrane that ultimately results in cell (NEWTON 1956; SCHINDLER and OSBORN 1979). Previous studies have shown that decreased binding of drug to LPS causes colistin resistance in S. typhimurium. Resistance mutations were mapped to two genes, pmrA and pmrB (ADAMS et al. 2009; MOSKOWITZ et al. 2004; ROLAND et al. 1993; VAARA et al. 1979). PmrA and PmrB constitute a two-component regulatory system regulating the of several downstream genes that can modify the LPS (GUNN et al. 1998; GUNN et al. 2000). These modifications make the LPS less negatively charged and thereby decrease the binding of colistin. In Paper V, we have determined the mutation rate to colistin resistance and examined the fitness of the resistant mutants in vitro and in a mouse model. Mutants with reduced susceptibility to colistin were isolated by plating ~2×107 cells from independent cultures of S. typhimurium onto LA plates supplemented with colistin. 44 spontaneous mutants with reduced susceptibility to colistin were purified and tested for resistance to colistin. The number of colonies from each plate was counted and the mutation rate was calculated to be about 6×10-7 per cell per generation by using the median method of Lea and Coulson (LEA D. E. 1949), which is several orders of magnitude higher than for many conventional antibiotics (POPE et al. 2008). The minimum inhibitory concentration (MIC) of colistin was determined for the mutants and the wild type strain, and for most mutants, the MIC was increased about 20- and 30-fold compared to the wild type strain. Killing assay showed that the mutants had a much higher ability to survive compared to the wild type strain when treated with colistin and the mutants with high MICs were more resistant to the killing effect. Sequencing of the pmrA and pmrB gene and genetic analysis confirmed that missense mutations in these two genes were responsible for the reduced susceptibility of mutants to colistin. The expression level of the pmrH gene, which is one of the genes regulated by PmrA-PmrB two-component system, was measured by real-time PCR for both mutants and the wild type strain. The increased expression levels of the pmrH gene in mutants compared to the wild type strain indicated that identified pmrA/pmrB mutations changed the regulation of this two-component system. A detailed examination of the locations of the mutations in the pmrB gene based on a homologous protein structure indicated that the mutations in the PmrB protein are likely

37 to increase the kinase/phosphatase activity ratio and in this way alter the regulation of the PmrA-PmrB regulatory system. The exponential growth rates of 5 pmrA mutants and 18 pmrB mutants were measured in LB medium and M9- without colistin. The fitness costs were higher in poor medium than in rich medium, and in both media, the pmrA mutants showed a more severe reduction in growth than the pmrB mutants. We also examined the potential of these mutants to survive prolonged starvation, no significant differences were observed between the four mutants and the wild type strain. Since the parameters affecting fitness can be very different in vitro compared to in vivo, we also determined the competitive ability of the mutants in a mouse infection model. No significant differences in fitness were detected between the mutants and the wild type strain. The relatively low (in vitro) or absent (in vivo) fitness cost of the pmrAB mutations and the high mutation rate suggest that mutants with reduced susceptibility to colistin could relatively easily be selected clinically, which is in accordance with the current clinical situation (ANTONIADOU et al. 2007; HAWLEY et al. 2008; KO et al. 2007; PITT et al. 2003; SCHÜLIN 2002; THIOLAS et al. 2005).

Evolution of a novel Metallo-ß-lactamase The most common mechanism of ß-lactam resistance is the production of ß- lactamases that can hydrolyze the ß-lactam ring present in all ß-lactam antibiotics (MAJIDUDDIN et al. 2002; MEDEIROS 1984). Based on their primary , ß-lactamases are classified into four major classes (A-D) (AMBLER 1980; JAURIN and GRUNDSTRÖM 1981; OUELLETTE et al. 1987). Class A, C and D enzymes employ a serine residue in their active sites, while class B enzymes, metallo-ß-lactamases (MBLs) use one or two divalent cations (Zn2+) to coordinate two water molecules as the reactive nucleophiles (FISHER et al. 2005; HELFAND and BONOMO 2003). MBLs are particularly problematic due to their ability to hydrolyze virtually all classes of ß-lactam antibiotics (WALSH et al. 2005). In paper VI, we evolved an engineered MBL towards higher resistance against seven different β-lactam antibiotics by and examined the functional consequences of this adaptive evolution process. In a recent study, a MBL activity was elicited on a metallohydrolase scaffold of human glyoxalase II (PARK et al. 2006). The resulting enzyme, evMBL8, completely lost its original function and instead was able to hydrolyze cefotaxime. A modified version of evMBL8 (designated as evMBL9) exhibited a small but reproducibly higher MIC against almost all tested β-lactam antibiotics. Therefore, evMBL9 was employed as the ancestral MBL in the evolution experiments. To efficiently explore sequence and generate the right diversity, we designed and synthesized a mutant

38 in which the substrate binding profile was varied by randomizing six amino acid residues, which were predicted to be important for substrate binding based on molecular modeling of evMBL8. In addition, two random mutagenesis libraries were constructed allowing us to examine the functionalities of all residues in the enzyme. After cloning the three libraries into S. typhimurium, mutants with increased resistance against seven different ß-lactam antibiotics (penicillin G, ampicillin, cefalotin, cefaclor, cefuroxime, and cefotaxime) were isolated and characterized. For the majority of mutants, in spite of their significantly increased resistance, both evMBL9 mRNA and protein levels were reduced, indicating that the catalytic activities of these mutant MBLs were highly increased. Cross-resistance to six other antibiotics was determined for mutants with increased resistance to the corresponding antibiotics used for selection and the resistance profiles of selected mutants were analyzed by multivariate analysis (KURTOVIC and MANNERVIK 2009). The results indicated that the majority of mutant enzymes became generalists, conferring increased resistance against most of the examined ß-lactams. The increased resistance and decreased protein level suggest that the improved hydrolysis in these novel MBLs is associated with decreased protein stability.

Connecting penicillin binding proteins and β-lactamases We have initiated a project to examine the potential for penicillin binding proteins to evolve towards β-lactamases. Penicillin binding proteins (PBPs) catalyze polymerization of the glycan strand and the cross-linking between glycan chains (SAUVAGE et al. 2008). The penicillin binding domains of PBPs function as transpeptidases or carboxypeptidases that are involved in (GOFFIN and GHUYSEN 1998; SAUVAGE et al. 2008). These domains have three signature motifs: SXXK, (S/Y)XN and (K/H)(S/T)G, which are shared by the class A and class C ß-lactamases (GHUYSEN 1991; ZAPUN et al. 2008). For both PBPs and ß-lactamases, the active serine residue attacks the ß-lactam ring forming a covalent acyl- enzyme complex. The deacylation step is very fast with ß-lactamases but extremely slow with PBPs (GHUYSEN 1991). It has been postulated that ß- lactamases evolved from penicillin binding proteins (MASSOVA and MOBASHERY 1998). Fifteen genes encoding defined or putative penicillin binding proteins in S. typhimurium genome (ampH, dacA, dacB, dacC, dacD, mrcA, mrcB, mrdA, pbpC, pbpG, STM1836, STM1910, STM2478, ftsI, and yfeL) were cloned in pUCBAD-kan plasmid and transformed into wild type S. typhimurium strain. The strains carrying PBPs on plasmids and the wild type strain with empty plasmid were grown overnight at 37°C in the presence of 2mM arabinose and MICs of penicillin G were determined. No significant MIC increase was observed for all the tested strains under

39 arabinose induction as compared with the control strain. We performed random mutagenesis on PBPs followed by penicillin G selection to examine the potential of these PBPs to increase their resistance against penicillin G. Based on the results from the first round of mutagenesis and selection, the evolution experiments were focused on the ftsI gene encoding PBP3 and the dacC gene encoding PBP6 because mutations in these two genes could confer 2-3 fold increases of MIC against penicillin-G. A second round of mutagenesis and selection was also performed to further evolve PBP3 and PBP6 towards higher resistance. In most cases, multiple mutations were identified in the mutant PBPs. At present, we are trying to dissect the potential effect of the identified mutations by reconstructing single mutations and determining the MIC values of penicillin G.

Concluding remarks Microorganisms have been present on for billions of years and are of great importance to humans, not only because many microbes are pathogens but also for the important roles they play in all types of . Therefore, it is crucial to understand the mechanisms of microbial evolution. Bacteria have the ability to rapidly adapt to any changing environment and have been successfully used in laboratory evolution experiments to unravel the mechanisms and principles of adaptation. In this thesis, I have performed several studies to investigate the dynamics and genetic basis of adaptive evolution in bacteria using S. typhimurium as the model organism. Gene amplification has been found in diverse groups of organisms (DUNHAM et al. 2002; EL-SAYED et al. 2005; GUILLEMAUD et al. 1999; ROMERO et al. 1995; WONG et al. 2007) and is important both from a fundamental evolutionary perspective (BERGTHORSSON et al. 2007; OHNO 1970) and in medical genetics as an important contributor to human diseases and phenotypic variability among individuals (BECKMANN et al. 2007; CONRAD and ANTONARAKIS 2007). In paper I, we examined the role of the amplification model in establishing new gene functions by formalizing it into a mathematical framework and performing stochastic simulations. The parameter values used in the simulations were based on the experimental measurements of the fitness costs and loss rates for a set of duplication- carrying S. typhimurium strains. The results demonstrate that gene amplification is likely to contribute to creation of new gene functions in nature. In paper II, we experimentally evolved a S. typhimurium strain towards increased resistance against cephalosporins. This strain carried a β- lactamase gene (blaTEM-1) with very low resistance to cephalosporins. The initial response was amplification of the blaTEM-1 gene. In some cases, amplification was followed by acquisition of mutations that reduced the expression of uptake functions. Our results suggest that gene amplification

40 can facilitate the adaptive evolution process by allowing the population to expand sufficiently to realize stable adaptive mutations. Deletions are generally thought to have a deleterious effect on organism fitness and the main driving force of reductive evolution in endosymbionts is believed to be increased genetic drift. In paper III, the alternative hypothesis that adaptive processes contribute to genome reduction was subject to experimental examination. In the bacterium S. typhimurium, we determined the rates and fitness effects of spontaneous deletions in different chromosomal positions. Approximately 25% of examined deletions had increased fitness levels in at least one of the three tested conditions. This was further supported by the observation that deletions were frequently fixed in wild type bacterial populations passaged for 1000 generations. These findings suggest that selection can drive genome reduction. As suggested in paper I, II and III, genome rearrangements such as duplications and deletions play important roles in bacterial adaptive evolution and thus it is fundamental to understand the dynamics of spontaneous genome rearrangements (SGRs) in unselected bacterial populations. Previous estimations suggest that at least 10% of the cells contain a duplication somewhere in the genome in a growing S. typhimurium population (ANDERSON and ROTH 1981). The frequency of deletions and inversions, have not previously been measured on a genome-wide scale. In Paper IV, we developed a new strategy combining 454 pyrosequencing technology and a ‘split mapping’ computational method to investigate dynamics of SGRs in a continuously growing bacterial population by detecting their unique junction sequences. Our results demonstrate the high steady-state frequency of rearrangements in bacterial populations. Among the many examples of evolution in action that have been studied in nature, the emergence of antibiotic resistance is probably the most significant one. In paper V, we focused our experimental evolution study on the development of bacterial resistance to an old antibiotic, colistin. Colistin was discovered in 1949 and was later abandoned due to its . The emergence of multidrug-resistant gram-negative bacteria and lack of effective antibiotics led to a revival of this old antibiotic. In this study, we measured the spontaneous mutation rate and fitness effects of colistin resistant mutants in S. typhimurium. The mutants appeared at a rate of 0.6×10-6 per cell per generation. Low fitness costs were observed for resistant mutants in vitro and growth in mice was unaffected by the resistance mutations. These results suggest that mutants with reduced susceptibility to colistin could easily emerge clinically. Proteins are key to most biological processes and can carry out biochemical functions with tremendous diversity. Therefore it is essential to understand the mechanisms of adaptive protein evolution. In paper VI, we employed an engineered metallo-ß-lactamase (MBL) as the ancestral protein and performed a directed evolution experiment to examine how an enzyme

41 evolved towards increased resistance. The main strategy used to generate protein variants was to randomize six amino acid positions that were predicted to be important for substrate binding and the desired target property is increased resistance against seven different β-lactam antibiotics. Our result suggest that at the initial stage of MBL evolution: (i) the evolved proteins increase their specific activity and are destabilized, and (ii) most evolved MBLs become ‘generalist’ enzymes that confer increased resistance against different types of β-lactam antibiotics.

42 Future perspectives

Based on the studies included in this thesis, several interesting follow-up projects have emerged:

In paper II, following the tandem amplification of the blaTEM-1 gene that was responsible for cephalosporin resistance at the initial stage, the subsequent stable mutations were identified in genes located on the chromosome but not in the β-lactamase gene blaTEM-1. This is likely to be due to the fact that the target size for mutations in chromosomal genes is much bigger than in the single blaTEM-1 gene. However, these mutations conferred resistance by the same mechanism, i.e. by reducing expression of porin proteins. Thus, it would be interesting to inactivate the chromosomal genes encoding the major porin proteins such as OmpC, OmpF and OmpD and repeat the evolution experiments to examine whether mutations in the blaTEM-1 gene would appear following the amplification. If TEM-1 mutations were selected and conferred resistance by improving the catalytic activity of hydrolyzing cephalosporins, do they impair the original TEM-1 function? Is there cross- resistance between different cephalosproins? The answers to these questions will not only provide insights into the roles of gene amplification in creation of new gene functions but also shed lights on the mechanisms of adaptive protein evolution. In paper III, we have been able to isolate deletion mutants that showed a selective advantage when competed with the isogenic wild type strains. It would be interesting to randomly combine these beneficial deletions and examine the potential epistatic interactions. Negative between beneficial mutations have been implicated in bacterial populations (KHAN et al. 2011). Since the beneficial deletions isolated in this work usually includes several genes, will they exhibit different types of epistasis or is diminishing return dominant in epistatic interactions between beneficial mutations? In this work, we have experimentally evolved six independent lineages of wild type S. typhimurium in rich LB medium for 1000 generations and six deletions were fixed in the population. Two out of the six deletion mutations increased fitness but the other four had reduced fitness when reconstructed in a wild type genetic background. It is likely that the deleterious deletions increased fitness in combination with other mutations and were also fixed by selection. One way to test this hypothesis is to reverse

43 the deletion mutation to wild type allele in the deletion mutant and examine the fitness effects of the reversion. In paper IV, we developed a new strategy to detect spontaneous genome rearrangements (SGRs) on a genome-wide scale based on in silico screening and experimental verification of rearrangement junction sequences from 454 pyrosequencing data. Using this approach, we were able to estimate the steady-state frequency of SGRs in an unselected bacterial population. One interesting project would be to compare the frequencies of SGRs in a wild type strain to strains with mutations (i.e., recA, mutS, mutL) predicted to affect the formation of SGRs, which will further increase our understanding of the genetic basis of SGRs. Similarly, we could also examine whether the steady-state frequency of SGRs is affected by different growth conditions. Large amounts of sequences from numerous microbial genome sequencing projects have been deposited in NCBI trace archive and it would be interesting to estimate the frequency of SGRs for microbial species from different phylogenetic groups and to examine the potential relationship between the frequency of SGRs and the genomic stability inferred from comparative genomic studies.

44 摘要

适应性进化通常会导致物种产生复杂的遗传特性， 但究其根本此过程只取决 于两个因素： 突变和选择。 基因突变是遗传变异的源泉，从而为自然选择提 供了原材料，因此确定基因突变的特性对于理解适应性进化的机制具有至关重 要的意义。这篇论文的主要目的是以实验性进化为手段，通过对不同类型突变 （特别是基因组重排）的研究，来阐释细菌适应性进化的机制。在细菌进化实 验中，所使用的模型生物是沙门氏杆菌，而选择压力则主要是对抗生素耐药性 的提高。 论文由六个相互关联而又各有侧重的课题组成。基因扩增对细菌的适应性 进化起着重要的作用。在第一个课题中，我们通过数学模型模拟了基因扩增的 过程，并结合实验进化的数据验证了基因扩增对于新基因功能建立的重要意义。 在第二个课题中，通过细菌进化实验，我们逐步提高了一个携带β-内酰胺酶的 沙门氏杆菌菌株对头孢菌素类抗生素的耐药性，研究结果表明基因扩增可以在 细菌适应性进化的早期直接起到缓解选择压力的作用，并进一步帮助细菌获得 稳定的适应性遗传性状。与基因扩增相反，缺失突变会导致基因组的缩小，而 这种缩减进化的主要驱动力被普遍认为是遗传漂移。在第三个课题中，我们分 离了自发突变缺失体菌株并发现大约 25％的突变菌株的适应性显著高于野生 型菌株；在随后的野生型菌株连续传代实验中，我们观察到缺失突变在种群中 被自然选择固定下来。这些发现证明了自然选择可以驱动缩减进化。在细菌中， 基因重排主要包括基因扩增，缺失和倒置。在第四个课题中，通过对 454 焦磷 酸测序技术和“分离映射” 算法的结合，我们开发出一种新的方法来识别由 自发基因重排所形成的接界序列。运用这种方法，我们对未经过选择的细菌种 群的自发基因重排进行了分析，结果表明自发基因重排在细菌种群种非常普遍。 细菌耐药性尤其是多抗性革兰氏阴性菌的大量出现使多粘菌素又重新回到人们 的视线中。在第五个课题中，我们研究了沙门氏杆菌对多粘菌素的耐药突变率， 耐药机理和耐药菌株的适应缺陷。较高的突变率和较低的适应性缺陷表明耐多 粘菌素菌株有可能在临床中出现。在第六个课题中，我们利用蛋白质定向进化 实验提高了一个新型金属β-内酰胺酶对β-内酰胺类抗生素的耐药性，分离并鉴 定了突变体蛋白的特性。尽管突变体金属β-内酰胺酶显著提高了细菌的耐药性， 他们的信使核糖核酸和蛋白表达水平确远低于野生型酶，这一结果表明突变体 金属β-内酰胺酶的催化活性大幅度提高。

关键词：适应性进化，突变，基因重排，抗生素耐药，基因扩增，缩减进 化，定向进化

45 Acknowledgements

It would not be possible to write this thesis without the support and help from the nice people around me and I would here like to express my sincere gratitude to some of them.

First of all, I am heartily thankful to my supervisor Dan Andersson, who has introduced me into the scientific world and guided me through my doctoral studies with his knowledge and wisdom. Dan, thank you for inspiring and encouraging me to become an independent researcher and being so supportive in every way. It was my great honor to pursue a PhD degree under your supervision.

I would like to extend my gratitude to my co-supervisor Diarmaid Hughes for all the helpful discussions and constructive comments on my research work.

My excellent coauthors , Otto Berg, Mats Pettersson, Aurel Negrea, Mikael Rhen, Wei Zhang, Bengt Mannervik, Rongqin Ke, Mats Nilsson, Diarmaid Hughes and Sanna Koskiniemi for various interesting discussions and for the fruitful collaborations.

Present and former members of the DA-lab. Linus, the rising star of science, thank you for being a great listener and for teaching me how to ski properly. Sanna, you have set a great example of growing up from a PhD student to an excellent young scientist. Your hard work and enthusiasm have motivated me a lot. Anna Z, I always admire your ability of analyzing and solving problems and great skills. Thank you for inspiring me in Perl programming. Maria, you are such a kind and helpful person. Your rigorous research attitude is something that I will bear firmly in mind as a researcher. Chris, the one with the most powerful laugher in the lab, your encouragement and kindness mean a lot to me. Thank you for the help on my thesis writing. Peter, the master of evolutionary theory, thanks for inspiring me with your brilliant scientific work. Jocke, the one who really enjoys doing research, thank you for always giving the right answers to my questions. Annika, thank you for helping me with my first evolution experiment. Herve, I am grateful for your constructive advice on my

46 research work and the joy you bring to the lab. Ulrika, thank you for all the kind help and I can’t imagine how our lab can operate without you. Amira, it has been a pleasure to get to know you. Marlen, now the most senior PhD student and the one who can always get things done in the lab. Thank you for being a perfect roommate. Lisa T, you bring such positivity to the lab and thanks for teaching me Swedish during coffee times although I am not a very good student. Hava, always so calm but also have a highly contagious smile. Erik, you are so smart and keep amazing me with your scientific work. Marius, you have a natural affinity to people around you and I think you are so far the best supporting actor in DA movies. To the new generation of PhD students in DA-lab, Anna K, Lisa A, and Michael, wish you best luck in the future and you will keep the DA-lab successful.

Special thanks to Maria and Jocke for comments and of this thesis.

I would like to thank all of my IMBIM colleagues especially the D7:3 corridor for creating a great working environment and all the IMBIM staff for being so helpful.

I would also like to thank all of those Chinese friends that have made my life here in Uppsala so wonderful. Firstly, the people from the badminton troop, for bringing me into the world of this exciting sport and more importantly for being great friends over the years; the people from the card game club for all those “serious” card plays starting at Saturday nights and ending (sometimes restarting) at Sunday mornings; those lunch fellows in BMC for taking the pressure off me and all the laid-back talks; the people from the basketball team for keeping me work hard in the gym; all the classmates in MBB2006 for the lectures, parties, parties, and parties.

To my parents, my mam Ping and my dad Weimin, 感谢爸爸妈妈对我的 养育之恩，谢谢你们一直包容我，支持我，鼓励我。

To my wife Na, I am so lucky to have you in my life and I wouldn’t have made it this far without your support and love. Thank you for giving me the best present of my life – our son Zubin.

47 References

ADAMS, M. D., G. C. , S. BAJAKSOUZIAN, H. LAVENDER, A. R. MURTHY et al., 2009 Resistance to colistin in associated with mutations in the PmrAB two-component system. Antimicrob Agents Chemother 53: 3628-3634. ALBERTINI, A. M., M. HOFER, M. P. CALOS and J. H. MILLER, 1982 On the formation of spontaneous deletions: the importance of short sequence homologies in the generation of large deletions. Cell 29: 319-328. ALBERTSON, D. G., 2006 Gene amplification in cancer. Trends Genet 22: 447-455. ALEKSHUN, M. N., and S. B. LEVY, 2007 Molecular mechanisms of antibacterial multidrug resistance. Cell 128: 1037-1050. ALOKAM, S., S. L. LIU, K. SAID and K. E. SANDERSON, 2002 Inversions over the terminus region in Salmonella and Escherichia coli: IS200s as the sites of homologous recombination inverting the chromosome of Salmonella enterica serovar typhi. J Bacteriol 184: 6190-6197. AMBLER, R. P., 1980 The structure of beta-lactamases. Philos Trans R Soc Lond, B, Biol Sci 289: 321-331. AMORES, A., A. FORCE, Y. L. YAN, L. JOLY, C. AMEMIYA et al., 1998 Zebrafish hox clusters and vertebrate . Science 282: 1711-1714. AMYES, S. G., 2003 Resistance to beta-lactams--the permutations. Journal of (Florence, Italy) 15: 525-535. ANDERSON, P., and J. ROTH, 1981 Spontaneous tandem genetic duplications in Salmonella typhimurium arise by unequal recombination between rRNA (rrn) cistrons. Proc Natl Acad Sci U S A 78: 3113-3117. ANDERSON, R. P., and J. R. ROTH, 1977 Tandem genetic duplications in phage and bacteria. Annu Rev Microbiol 31: 473-505. ANDERSON, R. P., and J. R. ROTH, 1978 Tandem chromosomal duplications in Salmonella typhimurium: fusion of histidine genes to novel promoters. J Mol Biol 119: 147-166. ANDERSSON, D. I., 2003 Persistence of antibiotic resistant bacteria. Curr Opin Microbiol 6: 452-456. ANDERSSON, D. I., 2006 The biological cost of mutational antibiotic resistance: any practical conclusions? Curr Opin Microbiol 9: 461- 465. ANDERSSON, D. I., 2011 Evolving promiscuously. Proc Natl Acad Sci U S A 108: 1199-1200.

48 ANDERSSON, D. I., and D. HUGHES, 2009 Gene amplification and adaptive evolution in bacteria. Annu Rev Genet 43: 167-195. ANDERSSON, D. I., and B. R. LEVIN, 1999 The biological cost of antibiotic resistance. Curr Opin Microbiol 2: 489-493. ANDERSSON, D. I., E. S. SLECHTA and J. R. ROTH, 1998a Evidence that gene amplification underlies adaptive mutability of the bacterial lac operon. Science 282: 1133-1135. ANDERSSON, J. O., and S. G. ANDERSSON, 1999 Genome degradation is an ongoing process in Rickettsia. Mol Biol Evol 16: 1178-1191. ANDERSSON, J. O., and S. G. ANDERSSON, 2001 Pseudogenes, junk DNA, and the dynamics of Rickettsia genomes. Mol Biol Evol 18: 829- 839. ANDERSSON, S. G., and C. G. KURLAND, 1998 Reductive evolution of resident genomes. Trends Microbiol 6: 263-268. ANDERSSON, S. G., A. ZOMORODIPOUR, J. O. ANDERSSON, T. SICHERITZ- PONTEN, U. C. ALSMARK et al., 1998b The genome sequence of and the origin of mitochondria. Nature 396: 133-140. ANTONIADOU, A., F. KONTOPIDOU, G. POULAKOU, E. KORATZANIS, I. GALANI et al., 2007 Colistin-resistant isolates of emerging in intensive care unit patients: first report of a multiclonal cluster. J Antimicrob Chemother 59: 786-790. ASHIZAWA, Y., T. YOKOCHI, Y. OGATA, Y. SHOBUIKE, J. KATO et al., 1999 Mechanism of DNA gyrase-mediated illegitimate recombination: characterization of Escherichia coli gyrA mutations that confer hyper-recombination phenotype. J Mol Biol 289: 447-458. BARTH, P. T., N. DATTA, R. W. HEDGES and N. J. GRINTER, 1976 Transposition of a deoxyribonucleic acid sequence encoding trimethoprim and streptomycin resistances from R483 to other replicons. J Bacteriol 125: 800-810. BECKMANN, J. S., X. ESTIVILL and S. E. ANTONARAKIS, 2007 Copy number variants and genetic traits: closer to the resolution of phenotypic to genotypic variability. Nat Rev Genet 8: 639-646. BENEDIK, M., M. FENNEWALD and J. SHAPIRO, 1977 Transposition of a beta-lactamase locus from RP1 into degradative plasmids. J Bacteriol 129: 809-814. BERGTHORSSON, U., D. I. ANDERSSON and J. R. ROTH, 2007 Ohno's dilemma: evolution of new genes under continuous selection. Proc Natl Acad Sci U S A 104: 17004-17009. BERGTHORSSON, U., and H. OCHMAN, 1995 Heterogeneity of genome sizes among natural isolates of Escherichia coli. J Bacteriol 177: 5784- 5789. BI, X., and L. F. LIU, 1994 recA-independent and recA-dependent intramolecular plasmid recombination. Differential homology requirement and distance effect. J Mol Biol 235: 414-423. BI, X., and L. F. LIU, 1996 DNA rearrangement mediated by inverted repeats. Proc Natl Acad Sci U S A 93: 819-823.

49 BIERNE, H., S. D. EHRLICH and B. MICHEL, 1997a Deletions at stalled replication forks occur by two different pathways. EMBO J 16: 3332-3340. BIERNE, H., M. SEIGNEUR, S. D. EHRLICH and B. MICHEL, 1997b uvrD mutations enhance deletion in the Escherichia coli chromosome via SOS induction of the RecF recombination pathway. Mol Microbiol 26: 557-567. BIERNE, H., D. VILETTE, S. D. EHRLICH and B. MICHEL, 1997c of a dnaE mutation which enhances RecA-independent homologous recombination in the Escherichia coli chromosome. Mol Microbiol 24: 1225-1234. BJORKMAN, J., D. HUGHES and D. I. ANDERSSON, 1998 of antibiotic-resistant Salmonella typhimurium. Proc Natl Acad Sci U S A 95: 3949-3953. BJÖRKMAN, J., I. NAGAEV, O. G. BERG, D. HUGHES and D. I. ANDERSSON, 2000 Effects of environment on compensatory mutations to ameliorate costs of antibiotic resistance. Science 287: 1479-1482. BOE, L., M. DANIELSEN, S. KNUDSEN, J. B. PETERSEN, J. MAYMANN et al., 2000 The frequency of mutators in populations of Escherichia coli. Mutat Res 448: 47-55. BOTTGER, E. C., B. SPRINGER, M. PLETSCHETTE and P. SANDER, 1998 Fitness of antibiotic-resistant microorganisms and compensatory mutations. Nat Med 4: 1343-1344. BREWER, B. J., 1988 When polymerases collide: replication and the transcriptional organization of the E. coli chromosome. Cell 53: 679-686. BROCHET, M., E. COUVE, M. ZOUINE, C. POYART and P. GLASER, 2008 A naturally occurring gene amplification leading to sulfonamide and trimethoprim resistance in . J Bacteriol 190: 672-680. BRUSSOW, H., C. CANCHAYA and W. D. HARDT, 2004 Phages and the evolution of bacterial pathogens: from genomic rearrangements to lysogenic conversion. Microbiol Mol Biol Rev 68: 560-602. BZYMEK, M., and S. T. LOVETT, 2001 Instability of repetitive DNA sequences: the role of replication in multiple mechanisms. Proc Natl Acad Sci U S A 98: 8319-8325. CAIRNS, J., and P. L. FOSTER, 1991 Adaptive reversion of a frameshift mutation in Escherichia coli. Genetics 128: 695-701. CAIRNS, J., J. OVERBAUGH and S. MILLER, 1988 The origin of mutants. Nature 335: 142-145. CAMPO, N., M. J. DIAS, M. L. DAVERAN-MINGOT, P. RITZENTHALER and P. LE BOURGEOIS, 2004 Chromosomal constraints in Gram-positive bacteria revealed by artificial inversions. Mol Microbiol 51: 511- 522. CASJENS, S., 1998 The diverse and dynamic structure of bacterial genomes. Annu Rev Genet 32: 339-377.

50 CHEN, I., P. J. CHRISTIE and D. DUBNAU, 2005 The ins and outs of DNA transfer in bacteria. Science 310: 1456-1460. CLARK, A. G., 1994 Invasion and maintenance of a gene duplication. Proc Natl Acad Sci U S A 91: 2950-2954. CLEWELL, D. B., Y. YAGI and B. BAUER, 1975 Plasmid-determined tetracycline resistance in Streptococcus faecalis: evidence for gene amplification during growth in presence of tetracycline. Proc Natl Acad Sci U S A 72: 1720-1724. COLE, S. T., K. EIGLMEIER, J. PARKHILL, K. D. JAMES, N. R. THOMSON et al., 2001 Massive gene decay in the bacillus. Nature 409: 1007-1011. CONRAD, B., and S. E. ANTONARAKIS, 2007 Gene duplication: a drive for phenotypic diversity and cause of human . Annu Rev Hum Genet 8: 17-35. CONRAD, T. M., N. E. LEWIS and B. O. PALSSON, 2011 Microbial laboratory evolution in the era of genome-scale science. Mol Syst Biol 7: 509. COOPER, V. S., and R. E. LENSKI, 2000 The of ecological specialization in evolving Escherichia coli populations. Nature 407: 736-739. CRAVEN, S. H., and E. L. NEIDLE, 2007 Double trouble: medical implications of genetic duplication and amplification in bacteria. Future 2: 309-321. DARLING, A. E., I. MIKLOS and M. A. RAGAN, 2008 Dynamics of genome rearrangement in bacterial populations. PLoS Genet 4: e1000128. DARWIN, C., 1859 On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. John Murray, London. DATTA, N., and P. KONTOMICHALOU, 1965 Penicillinase synthesis controlled by infectious R factors in Enterobacteriaceae. Nature 208: 239-241. DE VISSER, J. A., 2002 The fate of microbial mutators. Microbiology 148: 1247-1252. DENAMUR, E., and I. MATIC, 2006 Evolution of mutation rates in bacteria. Mol Microbiol 60: 820-827. DEVONSHIRE, A. L., and L. M. FIELD, 1991 Gene amplification and resistance. Annu Rev Entomol 36: 1-23. DRAKE, J. W., 1991 A constant rate of spontaneous mutation in DNA-based microbes. Proc Natl Acad Sci U S A 88: 7160-7164. DUNHAM, M. J., H. BADRANE, T. FEREA, J. ADAMS, P. O. BROWN et al., 2002 Characteristic genome rearrangements in experimental evolution of . Proc Natl Acad Sci U S A 99: 16144-16149. ECHOLS, H., and M. F. GOODMAN, 1991 Fidelity mechanisms in DNA replication. Annu Rev Biochem 60: 477-511. EDLUND, T., T. GRUNDSTROM and S. NORMARK, 1979 Isolation and characterization of DNA repetitions carrying the chromosomal beta-

51 lactamase gene of Escherichia coli K-12. Mol Gen Genet 173: 115- 125. EDLUND, T., and S. NORMARK, 1981 Recombination between short DNA homologies causes tandem duplication. Nature 292: 269-271. EGNER, C., and D. E. BERG, 1981 Excision of transposon Tn5 is dependent on the inverted repeats but not on the transposase function of Tn5. Proc Natl Acad Sci U S A 78: 459-463. EL-SAYED, N. M., P. J. MYLER, D. C. BARTHOLOMEU, D. NILSSON, G. AGGARWAL et al., 2005 The genome sequence of Trypanosoma cruzi, etiologic agent of Chagas disease. Science 309: 409-415. ENNE, V. I., A. A. DELSOL, G. R. DAVIS, S. L. HAYWARD, J. M. ROE et al., 2005 Assessment of the fitness impacts on Escherichia coli of acquisition of antibiotic resistance genes encoded by different types of genetic element. J Antimicrob Chemother 56: 544-551. ESNAULT, E., M. VALENS, O. ESPELI and F. BOCCARD, 2007 Chromosome structuring limits genome plasticity in Escherichia coli. PLoS Genet 3: e226. FERRIS, S. D., and G. S. WHITT, 1979 Evolution of the differential regulation of duplicate genes after polyploidization. J Mol Evol 12: 267-317. FISHER, J. F., S. O. MEROUEH and S. MOBASHERY, 2005 Bacterial resistance to beta-lactam antibiotics: compelling , compelling opportunity. Chem Rev 105: 395-424. FORCE, A., M. LYNCH, F. B. PICKETT, A. AMORES, Y. L. YAN et al., 1999 Preservation of duplicate genes by complementary, degenerative mutations. Genetics 151: 1531-1545. FOWLER, R. G., G. E. DEGNEN and E. C. COX, 1974 Mutational specificity of a conditional Escherichia coli mutator, mutD5. Mol Gen Genet 133: 179-191. FRASER, C. M., S. CASJENS, W. M. HUANG, G. G. SUTTON, R. CLAYTON et al., 1997 Genomic sequence of a Lyme disease , burgdorferi. Nature 390: 580-586. FRASER, C. M., J. D. GOCAYNE, O. WHITE, M. D. ADAMS, R. A. CLAYTON et al., 1995 The minimal gene complement of genitalium. Science 270: 397-403. FRIEDBERG, E. C., G. C. WALKER and W. SIEDE, 1995 DNA Repair and Mutagenesis. ASM Press, Washington, DC. FROST, L. S., R. LEPLAE, A. O. SUMMERS and A. TOUSSAINT, 2005 Mobile genetic elements: the agents of open source evolution. Nat Rev Microbiol 3: 722-732. GALITSKI, T., and J. R. ROTH, 1997 Pathways for homologous recombination between chromosomal direct repeats in Salmonella typhimurium. Genetics 146: 751-767. GHUYSEN, J. M., 1991 Serine beta-lactamases and penicillin-binding proteins. Annu Rev Microbiol 45: 37-67. GIRAUD, A., I. MATIC, M. RADMAN, M. FONS and F. TADDEI, 2002 Mutator bacteria as a risk factor in treatment of infectious diseases. Antimicrob Agents Chemother 46: 863-865.

52 GLICKMAN, B. W., and L. S. RIPLEY, 1984 Structural intermediates of deletion mutagenesis: a role for palindromic DNA. Proc Natl Acad Sci U S A 81: 512-516. GOFFIN, C., and J. M. GHUYSEN, 1998 Multimodular penicillin-binding proteins: an enigmatic family of orthologs and paralogs. Microbiol Mol Biol Rev 62: 1079-1093. GORDENIN, D. A., A. L. MALKOVA, A. PETERZEN, V. N. KULIKOV, Y. I. PAVLOV et al., 1992 Transposon Tn5 excision in : influence of DNA polymerases alpha, delta, and epsilon and repair genes. Proc Natl Acad Sci U S A 89: 3785-3789. GROSS, M. D., and E. C. SIEGEL, 1981 Incidence of mutator strains in Escherichia coli and coliforms in nature. Mutat Res 91: 107-110. GUILLEMAUD, T., M. RAYMOND, A. TSAGKARAKOU, C. BERNARD, P. ROCHARD et al., 1999 Quantitative variation and selection of esterase gene amplification in Culex pipiens. Heredity 83 ( Pt 1): 87-99. GUNN, J. S., K. B. LIM, J. KRUEGER, K. KIM, L. GUO et al., 1998 PmrA- PmrB-regulated genes necessary for 4-aminoarabinose lipid A modification and polymyxin resistance. Mol Microbiol 27: 1171- 1182. GUNN, J. S., S. S. RYAN, J. C. VAN VELKINBURGH, R. K. ERNST and S. I. MILLER, 2000 Genetic and functional analysis of a PmrA-PmrB- regulated locus necessary for lipopolysaccharide modification, resistance, and oral virulence of Salmonella enterica serovar typhimurium. Infect Immun 68: 6139-6146. HAACK, K. R., and J. R. ROTH, 1995 Recombination between chromosomal IS200 elements supports frequent duplication formation in Salmonella typhimurium. Genetics 141: 1245-1252. HALDANE, J. B. S., 1933 The Part Played by Recurrent Mutation in Evolution. The American Naturalist 67: 5-19. HALL, R. M., D. E. BROOKES and H. W. STOKES, 1991 Site-specific insertion of genes into integrons: role of the 59-base element and determination of the recombination cross-over point. Mol Microbiol 5: 1941-1959. HAMMOND, D. S., J. M. SCHOONEVELDT, G. R. NIMMO, F. HUYGENS and P. M. GIFFARD, 2005 bla(SHV) Genes in Klebsiella pneumoniae: different allele distributions are associated with different promoters within individual isolates. Antimicrob Agents Chemother 49: 256- 263. HASHIMOTO, H., and R. H. ROWND, 1975 Transition of the R factor NR1 and Proteus mirabilis: level of of nontransitioned and transitioned cells. J Bacteriol 123: 56-68. HASTINGS, P. J., H. J. BULL, J. R. KLUMP and S. M. ROSENBERG, 2000 Adaptive amplification: an inducible chromosomal instability mechanism. Cell 103: 723-731.

53 HAWLEY, J. S., C. K. MURRAY and J. H. JORGENSEN, 2008 Colistin heteroresistance in acinetobacter and its association with previous colistin therapy. Antimicrob Agents Chemother 52: 351-352. HELFAND, M. S., and R. A. BONOMO, 2003 Beta-lactamases: a survey of protein diversity. Current drug targets Infectious disorders 3: 9-23. HILL, C. W., J. FOULDS, L. SOLL and P. BERG, 1969 Instability of a missense suppressor resulting from a duplication of genetic material. J Mol Biol 39: 563-581. HILL, C. W., and J. A. GRAY, 1988 Effects of chromosomal inversion on cell fitness in Escherichia coli K-12. Genetics 119: 771-778. HILL, C. W., and B. W. HARNISH, 1981 Inversions between ribosomal RNA genes of Escherichia coli. Proc Natl Acad Sci U S A 78: 7069-7072. HILL, C. W., S. HARVEY and J. A. GRAY, 1990 Recombination between rRNA genes in Escherichia coli and Salmonella typhimurium. In The Bacterial Chromosome, ed. K Drlica, M Riley, pp. 335-40. Washington DC: ASM press. HOHMANN, E. L., 2001 Nontyphoidal . Clin Infect Dis 32: 263- 269. HUDSON, R. E., U. BERGTHORSSON, J. R. ROTH and H. OCHMAN, 2002 Effect of chromosome location on bacterial mutation rates. Mol Biol Evol 19: 85-92. HUYNEN, M. A., and P. BORK, 1998 Measuring genome evolution. Proc Natl Acad Sci U S A 95: 5849-5856. IKEDA, H., K. AOKI and A. NAITO, 1982 Illegitimate recombination mediated in vitro by DNA gyrase of Escherichia coli: structure of recombinant DNA molecules. Proc Natl Acad Sci U S A 79: 3724- 3728. IKEDA, H., K. SHIRAISHI and Y. OGATA, 2004 Illegitimate recombination mediated by double-strand break and end-joining in Escherichia coli. Adv Biophys 38: 3-20. IMHOF, M., and C. SCHLOTTERER, 2001 Fitness effects of advantageous mutations in evolving Escherichia coli populations. Proc Natl Acad Sci U S A 98: 1113-1117. JAURIN, B., and T. GRUNDSTRÖM, 1981 ampC cephalosporinase of Escherichia coli K-12 has a different evolutionary origin from that of beta-lactamases of the penicillinase type. Proc Natl Acad Sci USA 78: 4897-4901. JESSOP, A. P., and C. CLUGSTON, 1985 Amplification of the ArgF region in strain HfrP4X of E. coli K-12. Mol Gen Genet 201: 347-350. JONES, I. M., S. B. PRIMROSE and S. D. EHRLICH, 1982 Recombination between short direct repeats in a recA host. Mol Gen Genet 188: 486-489. JORDAN, I. K., K. S. MAKAROVA, J. L. SPOUGE, Y. I. WOLF and E. V. KOONIN, 2001 -specific gene expansions in bacterial and archaeal genomes. Genome Res 11: 555-565. JYSSUM, K., 1960 Observations on two types of genetic instability in Escherichia coli. Acta Pathol Microbiol Scand 48: 113-120.

54 KATO-MAEDA, M., J. T. RHEE, T. R. GINGERAS, H. SALAMON, J. DRENKOW et al., 2001 Comparing genomes within the species . Genome Res 11: 547-554. KHAN, A. I., D. M. DINH, D. SCHNEIDER, R. E. LENSKI and T. F. COOPER, 2011 Negative epistasis between beneficial mutations in an evolving bacterial population. Science 332: 1193-1196. KIBOTA, T. T., and M. LYNCH, 1996 Estimate of the genomic mutation rate deleterious to overall fitness in E. coli. Nature 381: 694-696. KIHARA, M., G. U. MILLER and R. M. MACNAB, 2000 Deletion analysis of the flagellar switch protein FliG of Salmonella. J Bacteriol 182: 3022-3028. KIMURA, M., 1983 The Neutral Theory of (Cambridge Univ. Press, Cambridge, 1983). KO, K. S., J. Y. SUH, K. T. KWON, S. I. JUNG, K. H. PARK et al., 2007 High rates of resistance to colistin and polymyxin B in subgroups of Acinetobacter baumannii isolates from Korea. J Antimicrob Chemother 60: 1163-1167. KONG, D., and W. MASKER, 1994 Deletion between direct repeats in T7 DNA stimulated by double-strand breaks. J Bacteriol 176: 5904- 5911. KONRAD, E. B., 1977 Method for the isolation of Escherichia coli mutants with enhanced recombination between chromosomal duplications. J Bacteriol 130: 167-172. KOWALCZYKOWSKI, S. C., 2000 Initiation of and recombination-dependent replication. Trends Biochem Sci 25: 156- 165. KRAWIEC, S., and M. RILEY, 1990 Organization of the bacterial chromosome. Microbiol Rev 54: 502-539. KUGELBERG, E., E. KOFOID, A. B. REAMS, D. I. ANDERSSON and J. R. ROTH, 2006 Multiple pathways of selected gene amplification during adaptive mutation. Proc Natl Acad Sci U S A 103: 17319- 17324. KUMARASAMY, K. K., M. A. TOLEMAN, T. R. WALSH, J. BAGARIA, F. BUTT et al., 2010 Emergence of a new antibiotic resistance mechanism in India, Pakistan, and the UK: a molecular, biological, and epidemiological study. Lancet Infect Dis 10: 597-602. KURTOVIC, S., and B. MANNERVIK, 2009 Identification of emerging quasi- species in directed enzyme evolution. 48: 9330-9339. KUTSUKAKE, K., H. NAKASHIMA, A. TOMINAGA and T. ABO, 2006 Two DNA invertases contribute to flagellar phase variation in Salmonella enterica serovar Typhimurium strain LT2. J Bacteriol 188: 950-957. LANG, A. S., and J. T. BEATTY, 2007 Importance of widespread genes in alpha-. Trends Microbiol 15: 54-62. LEA D. E., C., C. A. , 1949 The distribution of the numbers of mutants in bacterial populations. Journal of Genetics 49: 264-285.

55 LEHNER, A. F., and C. W. HILL, 1980 Involvement of ribosomal ribonucleic acid operons in Salmonella typhimurium chromosomal rearrangements. J Bacteriol 143: 492-498. LEVY, S. B., and B. MARSHALL, 2004 Antibacterial resistance worldwide: causes, challenges and responses. Nat Med 10: S122-129. LI, J., R. L. NATION, R. W. MILNE, J. D. TURNIDGE and K. COULTHARD, 2005 Evaluation of colistin as an agent against multi-resistant Gram- negative bacteria. Int J Antimicrob Agents 25: 11-25. LI, J., R. L. NATION, J. D. TURNIDGE, R. W. MILNE, K. COULTHARD et al., 2006 Colistin: the re-emerging antibiotic for multidrug-resistant Gram-negative bacterial infections. The Lancet infectious diseases 6: 589-601. LIANG, H., K. R. PLAZONIC, J. CHEN, W. H. LI and A. FERNANDEZ, 2008 Protein under-wrapping causes dosage sensitivity and decreases gene duplicability. PLoS Genet 4: e11. LIN, R. J., M. CAPAGE and C. W. HILL, 1984 A repetitive DNA sequence, rhs, responsible for duplications within the Escherichia coli K-12 chromosome. J Mol Biol 177: 1-18. LINDAHL, T., and R. D. WOOD, 1999 Quality control by DNA repair. Science 286: 1897-1905. LIU, G. R., W. Q. LIU, R. N. JOHNSTON, K. E. SANDERSON, S. X. LI et al., 2006 Genome plasticity and ori-ter rebalancing in Salmonella typhi. Mol Biol Evol 23: 365-371. LIU, S. L., and K. E. SANDERSON, 1996 Highly plastic chromosomal organization in Salmonella typhi. Proc Natl Acad Sci U S A 93: 10303-10308. LIVERMORE, D. M., 1995 beta-Lactamases in laboratory and clinical resistance. Clin Microbiol Rev 8: 557-584. LIVERMORE, D. M., 2004 The need for new antibiotics. Clin Microbiol Infect 10 Suppl 4: 1-9. LOVETT, S. T., P. T. DRAPKIN, V. A. SUTERA, JR. and T. J. GLUCKMAN- PESKIND, 1993 A sister-strand exchange mechanism for recA- independent deletion of repeated DNA sequences in Escherichia coli. Genetics 135: 631-642. LOVETT, S. T., T. J. GLUCKMAN, P. J. SIMON, V. A. SUTERA, JR. and P. T. DRAPKIN, 1994 Recombination between repeats in Escherichia coli by a recA-independent, proximity-sensitive mechanism. Mol Gen Genet 245: 294-300. LUPSKI, J. R., 1987 Molecular mechanisms for transposition of drug- resistance genes and other movable genetic elements. Rev Infect Dis 9: 357-368. LYNCH, M., and J. S. CONERY, 2000 The evolutionary fate and consequences of duplicate genes. Science 290: 1151-1155. LYNCH, M., and A. FORCE, 2000 The probability of duplicate gene preservation by subfunctionalization. Genetics 154: 459-473.

56 MACKAY, W. J., S. HAN and L. D. SAMSON, 1994 DNA alkylation repair limits spontaneous base substitution mutations in Escherichia coli. J Bacteriol 176: 3224-3230. MAHAN, M. J., and J. R. ROTH, 1988 Reciprocality of recombination events that rearrange the chromosome. Genetics 120: 23-35. MAISNIER-PATIN, S., and D. I. ANDERSSON, 2004 Adaptation to the deleterious effects of antimicrobial drug resistance mutations by compensatory evolution. Res Microbiol 155: 360-369. MAISNIER-PATIN, S., O. G. BERG, L. LILJAS and D. I. ANDERSSON, 2002 Compensatory adaptation to the deleterious effect of antibiotic resistance in Salmonella typhimurium. Mol Microbiol 46: 355-366. MAJIDUDDIN, F. K., I. C. MATERON and T. G. PALZKILL, 2002 Molecular analysis of beta-lactamase structure and function. Int J Med Microbiol 292: 127-137. MAKI, H., 2002 Origins of spontaneous mutations: specificity and directionality of base-substitution, frameshift, and sequence- substitution mutageneses. Annu Rev Genet 36: 279-303. MAO, E. F., L. LANE, J. LEE and J. H. MILLER, 1997 Proliferation of mutators in A cell population. J Bacteriol 179: 417-422. MARTÍNEZ-MARTÍNEZ, L., 2008 Extended-spectrum beta-lactamases and the permeability barrier. Clin Microbiol Infect 14 Suppl 1: 82-89. MASSOVA, I., and S. MOBASHERY, 1998 Kinship and diversification of bacterial penicillin-binding proteins and beta-lactamases. Antimicrob Agents Chemother 42: 1-17. MASTROENI, P., and M. SHEPPARD, 2004 Salmonella infections in the mouse model: host resistance factors and in vivo dynamics of bacterial spread and distribution in the tissues. Microbes Infect 6: 398-405. MATAGNE, A., A. M. MISSELYN-BAUDUIN, B. JORIS, T. ERPICUM, B. GRANIER et al., 1990 The diversity of the catalytic properties of class A beta-lactamases. Biochem J 265: 131-146. MATFIELD, M., R. BADAWI and W. J. BRAMMAR, 1985 Rec-dependent and Rec-independent recombination of plasmid-borne duplications in Escherichia coli K12. Mol Gen Genet 199: 518-523. MATTHEWS, P. R., and P. R. STEWART, 1988 Amplification of a section of chromosomal DNA in methicillin-resistant Staphylococcus aureus following growth in high concentrations of methicillin. J Gen Microbiol 134: 1455-1464. MCBETH, D. L., and J. A. SHAPIRO, 1984 Reversal by DNA amplifications of an unusual mutation blocking alkane and alcohol utilization in Pseudomonas putida. Mol Gen Genet 197: 384-391. MCCLELLAND, M., K. E. SANDERSON, J. SPIETH, S. W. CLIFTON, P. LATREILLE et al., 2001 Complete genome sequence of Salmonella enterica serovar Typhimurium LT2. Nature 413: 852-856. MCCUTCHEON, J. P., and N. A. MORAN, 2012 Extreme genome reduction in symbiotic bacteria. Nat Rev Microbiol 10: 13-26. MEDEIROS, A. A., 1984 Beta-lactamases. Br Med Bull 40: 18-27.

57 MIESEL, L., A. SEGALL and J. R. ROTH, 1994 Construction of chromosomal rearrangements in Salmonella by transduction: inversions of non- permissive segments are not lethal. Genetics 137: 919-932. MILLER, J. H., 1996 Spontaneous mutators in bacteria: insights into pathways of mutagenesis and repair. Annu Rev Microbiol 50: 625- 643. MIRA, A., H. OCHMAN and N. A. MORAN, 2001 Deletional bias and the evolution of bacterial genomes. Trends Genet 17: 589-596. MIZUUCHI, K., and T. BAKER, 2002 in Mobile DNA II. ASM press, Washington DC. MODRICH, P., 1991 Mechanisms and biological effects of mismatch repair. Annu Rev Genet 25: 229-253. MORAN, N. A., 1996 Accelerated evolution and Muller's rachet in endosymbiotic bacteria. Proc Natl Acad Sci U S A 93: 2873-2878. MORAN, N. A., 2002 Microbial minimalism: genome reduction in bacterial pathogens. Cell 108: 583-586. MORAN, N. A., 2003 Tracing the evolution of gene loss in obligate bacterial symbionts. Curr Opin Microbiol 6: 512-518. MORAN, N. A., and G. R. PLAGUE, 2004 Genomic changes following host restriction in bacteria. Curr Opin Genet Dev 14: 627-633. MOSKOWITZ, S. M., R. K. ERNST and S. I. MILLER, 2004 PmrAB, a two- component regulatory system of that modulates resistance to cationic antimicrobial and addition of aminoarabinose to lipid A. J Bacteriol 186: 575-579. MOYA, A., J. PERETO, R. GIL and A. LATORRE, 2008 Learning how to live together: genomic insights into - symbioses. Nat Rev Genet 9: 218-229. NADEAU, J. H., and D. SANKOFF, 1997 Comparable rates of gene loss and functional divergence after genome duplications early in vertebrate evolution. Genetics 147: 1259-1266. NAKABACHI, A., A. YAMASHITA, H. TOH, H. ISHIKAWA, H. E. DUNBAR et al., 2006 The 160-kilobase genome of the bacterial Carsonella. Science 314: 267. NEUBERGER, M. S., and B. S. HARTLEY, 1981 Structure of an experimentally evolved gene duplication encoding ribitol dehydrogenase in a mutant of . J Gen Microbiol 122: 181-191. NEWTON, B. A., 1956 The properties and mode of action of the polymyxins. Bacteriological reviews 20: 14-27. NICOLOFF, H., V. PERRETEN, L. M. MCMURRY and S. B. LEVY, 2006 Role for tandem duplication and lon protease in AcrAB-TolC- dependent multiple antibiotic resistance (Mar) in an Escherichia coli mutant without mutations in marRAB or acrRAB. J Bacteriol 188: 4413- 4423. NIKAIDO, H., 2003 Molecular basis of permeability revisited. Microbiol Mol Biol Rev 67: 593-656.

58 NILSSON, A. I., A. ZORZET, A. KANTH, S. DAHLSTROM, O. G. BERG et al., 2006 Reducing the fitness cost of antibiotic resistance by amplification of initiator tRNA genes. Proc Natl Acad Sci U S A 103: 6976-6981. NILSSON, M., H. MALMGREN, M. SAMIOTAKI, M. KWIATKOWSKI, B. P. CHOWDHARY et al., 1994 Padlock probes: circularizing oligonucleotides for localized DNA detection. Science 265: 2085- 2088. NORMARK, B. H., and S. NORMARK, 2002 Evolution and spread of antibiotic resistance. J Intern Med 252: 91-106. NOWAK, M. A., M. C. BOERLIJST, J. COOKE and J. M. SMITH, 1997 Evolution of genetic redundancy. Nature 388: 167-171. O'NEILL, A. J., T. HUOVINEN, C. W. FISHWICK and I. CHOPRA, 2006 Molecular genetic and structural modeling studies of Staphylococcus aureus RNA polymerase and the fitness of rifampin resistance in relation to clinical prevalence. Antimicrob Agents Chemother 50: 298-309. OCHMAN, H., and N. A. MORAN, 2001 Genes lost and genes found: evolution of bacterial pathogenesis and . Science 292: 1096-1099. OCHMAN, H., and A. C. WILSON, 1987 Evolution in bacteria: evidence for a universal substitution rate in cellular genomes. J Mol Evol 26: 74- 86. OGAWA, N., and K. MIYASHITA, 1995 Recombination of a 3-chlorobenzoate catabolic plasmid from eutrophus NH9 mediated by direct repeat elements. Appl Environ Microbiol 61: 3788-3795. OHNO, S., 1970 Evolution by gene duplication. Springer-Verlag, New York. OPPEZZO, O. J., B. AVANZATI and D. N. ANTÓN, 1991 Increased susceptibility to beta-lactam antibiotics and decreased porin content caused by envB mutations of Salmonella typhimurium. Antimicrob Agents Chemother 35: 1203-1207. OUELLETTE, M., L. BISSONNETTE and P. H. ROY, 1987 Precise insertion of antibiotic resistance determinants into Tn21-like transposons: nucleotide sequence of the OXA-1 beta-lactamase gene. Proc Natl Acad Sci USA 84: 7378-7382. PAPP, B., C. PAL and L. D. HURST, 2003 Dosage sensitivity and the evolution of gene families in yeast. Nature 424: 194-197. PARK, H. S., S. H. NAM, J. K. LEE, C. N. YOON, B. MANNERVIK et al., 2006 Design and evolution of new catalytic activity with an existing protein scaffold. Science 311: 535-538. PATRICK, W. M., E. M. QUANDT, D. B. SWARTZLANDER and I. MATSUMURA, 2007 Multicopy suppression underpins metabolic evolvability. Mol Biol Evol 24: 2716-2722. PAULANDER, W., D. I. ANDERSSON and S. MAISNIER-PATIN, 2010 Amplification of the gene for isoleucyl-tRNA synthetase facilitates adaptation to the fitness cost of mupirocin resistance in Salmonella enterica. Genetics 185: 305-312.

59 PERLMAN, D., and R. STICKGOLD, 1977 Selective amplification of genes on the R plasmid, NR1, in Proteus mirabilis: an example of the induction of selective gene amplification. Proc Natl Acad Sci U S A 74: 2518-2522. PETERSON, B. C., and R. H. ROWND, 1985 Drug resistance gene amplification of plasmid NR1 derivatives with various amounts of resistance determinant DNA. J Bacteriol 161: 1042-1048. PETES, T. D., and C. W. HILL, 1988 Recombination between repeated genes in microorganisms. Annu Rev Genet 22: 147-168. PETIT, M. A., J. M. MESAS, P. NOIROT, F. MOREL-DEVILLE and S. D. EHRLICH, 1992 Induction of DNA amplification in the chromosome. EMBO J 11: 1317-1326. PETROSINO, J., C. CANTU and T. PALZKILL, 1998 beta-Lactamases: protein evolution in real time. Trends Microbiol 6: 323-327. PITT, T. L., M. SPARROW, M. WARNER and M. STEFANIDOU, 2003 Survey of resistance of Pseudomonas aeruginosa from UK patients with to six commonly prescribed antimicrobial agents. Thorax 58: 794-796. POOLE, K., 2004 Resistance to beta-lactam antibiotics. Cell Mol Life Sci 61: 2200-2223. POPA, O., and T. DAGAN, 2011 Trends and barriers to lateral gene transfer in . Curr Opin Microbiol 14: 615-623. POPE, C. F., D. M. O'SULLIVAN, T. D. MCHUGH and S. H. GILLESPIE, 2008 A practical guide to measuring mutation rates in antibiotic resistance. Antimicrob Agents Chemother 52: 1209-1214. PORTNOY, V. A., D. BEZDAN and K. ZENGLER, 2011 Adaptive laboratory evolution--harnessing the power of biology for metabolic engineering. Curr Opin Biotechnol 22: 590-594. PRANTING, M., and D. I. ANDERSSON, 2011 Escape from growth restriction in small colony variants of Salmonella typhimurium by gene amplification and mutation. Mol Microbiol 79: 305-315. RANGNEKAR, V. M., 1988 Variation in the ability of Pseudomonas sp. strain B13 cultures to utilize meta-chlorobenzoate is associated with tandem amplification and deamplification of DNA. J Bacteriol 170: 1907-1912. REAMS, A. B., E. KOFOID, M. SAVAGEAU and J. R. ROTH, 2010 Duplication frequency in a population of Salmonella enterica rapidly approaches steady state with or without recombination. Genetics 184: 1077- 1094. REAMS, A. B., and E. L. NEIDLE, 2003 Genome plasticity in Acinetobacter: new degradative capabilities acquired by the spontaneous amplification of large chromosomal segments. Mol Microbiol 47: 1291-1304. REAMS, A. B., and E. L. NEIDLE, 2004 Gene amplification involves site- specific short homology-independent illegitimate recombination in Acinetobacter sp. strain ADP1. J Mol Biol 338: 643-656.

60 REBOLLO, J. E., V. FRANCOIS and J. M. LOUARN, 1988 Detection and possible role of two large nondivisible zones on the Escherichia coli chromosome. Proc Natl Acad Sci U S A 85: 9391-9395. RILEY, M., and A. ANILIONIS, 1978 Evolution of the bacterial genome. Annu Rev Microbiol 32: 519-560. ROLAND, K. L., L. E. MARTIN, C. R. ESTHER and J. K. SPITZNAGEL, 1993 Spontaneous pmrA mutants of Salmonella typhimurium LT2 define a new two-component regulatory system with a possible role in virulence. J Bacteriol 175: 4154-4164. ROMERO, D., J. MARTÍNEZ-SALAZAR, L. GIRARD, S. BROM, G. DÁVILLA et al., 1995 Discrete amplifiable regions (amplicons) in the symbiotic plasmid of etli CFN42. J Bacteriol 177: 973-980. ROMERO, D., and R. PALACIOS, 1997 Gene amplification and genomic plasticity in prokaryotes. Annu Rev Genet 31: 91-111. ROSCHE, W. A., and P. L. FOSTER, 2000 Determining mutation rates in bacterial populations. Methods 20: 4-17. ROTH, J. R., N. BENSON, T. GALITSKI, K. HAACK, J. G. LAWRENCE et al., 1996 Rearrangements of the bacterial chromosome: formation and applications. In Escherichia coli and Salmonella: Cellular and , 2nd ed., ed. FC Neidhardt, R Curtiss III, JL Ingraham, ECC Lin, KB Low, et al., pp. 2256-76. Washington D. C.: American Society for Microbiology. ROTH, J. R., E. KUGELBERG, A. B. REAMS, E. KOFOID and D. I. ANDERSSON, 2006 Origin of mutations under selection: the adaptive mutation controversy. Annu Rev Microbiol 60: 477-501. SALYERS, A. A., N. B. SHOEMAKER, A. M. STEVENS and L. Y. LI, 1995 Conjugative transposons: an unusual and diverse set of integrated gene transfer elements. Microbiol Rev 59: 579-590. SANDEGREN, L., and D. I. ANDERSSON, 2009 Bacterial gene amplification: implications for development of antibiotic resistance. Nat Rev Microbiol In press. SAUVAGE, E., F. KERFF, M. TERRAK, J. A. AYALA and P. CHARLIER, 2008 The penicillin-binding proteins: structure and role in peptidoglycan biosynthesis. FEMS Microbiol Rev 32: 234-258. SAVESON, C. J., and S. T. LOVETT, 1997 Enhanced deletion formation by aberrant DNA replication in Escherichia coli. Genetics 146: 457- 470. SCHAAPER, R. M., 1998 Antimutator mutants in T4 and Escherichia coli. Genetics 148: 1579-1585. SCHINDLER, M., and M. J. OSBORN, 1979 Interaction of divalent cations and polymyxin B with lipopolysaccharide. Biochemistry 18: 4425-4430. SCHMID, M. B., and J. R. ROTH, 1987 Gene location affects expression level in Salmonella typhimurium. J Bacteriol 169: 2872-2875. SCHNEIKER, S., O. PERLOVA, O. KAISER, K. GERTH, A. ALICI et al., 2007 Complete genome sequence of the myxobacterium . Nat Biotechnol 25: 1281-1289.

61 SCHÜLIN, T., 2002 In vitro activity of the aerosolized agents colistin and and five intravenous agents against Pseudomonas aeruginosa isolated from cystic fibrosis patients in southwestern Germany. J Antimicrob Chemother 49: 403-406. SEGALL, A., M. J. MAHAN and J. R. ROTH, 1988 Rearrangement of the bacterial chromosome: forbidden inversions. Science 241: 1314- 1318. SEGALL, A. M., and J. R. ROTH, 1989 Recombination between homologies in direct and inverse orientation in the chromosome of Salmonella: intervals which are nonpermissive for inversion formation. Genetics 122: 737-747. SEGALL, A. M., and J. R. ROTH, 1994 Approaches to half-tetrad analysis in bacteria: recombination between repeated, inverse-order chromosomal sequences. Genetics 136: 27-39. SHIGENOBU, S., H. WATANABE, M. HATTORI, Y. SAKAKI and H. ISHIKAWA, 2000 Genome sequence of the endocellular bacterial symbiont of Buchnera sp. APS. Nature 407: 81-86. SHIMIZU, H., H. YAMAGUCHI, Y. ASHIZAWA, Y. KOHNO, M. ASAMI et al., 1997 Short-homology-independent illegitimate recombination in Escherichia coli: distinct mechanism from short-homology- dependent illegitimate recombination. J Mol Biol 266: 297-305. SHIMIZU, H., H. YAMAGUCHI and H. IKEDA, 1995 Molecular analysis of lambda bio transducing phage produced by oxolinic acid-induced illegitimate recombination in vivo. Genetics 140: 889-896. SHIRAISHI, K., Y. IMAI, S. YOSHIZAKI and H. IKEDA, 2005 Rep helicase suppresses short-homology-dependent illegitimate recombination in Escherichia coli. Genes Cells 10: 1015-1023. SHYAMALA, V., E. SCHNEIDER and G. F. AMES, 1990 Tandem chromosomal duplications: role of REP sequences in the recombination event at the join-point. EMBO J 9: 939-946. SLECHTA, E. S., K. L. BUNNY, E. KUGELBERG, E. KOFOID, D. I. ANDERSSON et al., 2003 Adaptive mutation: general mutagenesis is not a programmed response to stress but results from rare coamplification of dinB with lac. Proc Natl Acad Sci U S A 100: 12847-12852. SLECHTA, E. S., J. LIU, D. I. ANDERSSON and J. R. ROTH, 2002 Evidence that selected amplification of a bacterial lac frameshift allele stimulates Lac(+) reversion (adaptive mutation) with or without general hypermutability. Genetics 161: 945-956. SNIEGOWSKI, P. D., P. J. GERRISH, T. JOHNSON and A. SHAVER, 2000 The evolution of mutation rates: separating causes from consequences. Bioessays 22: 1057-1066. SNIEGOWSKI, P. D., P. J. GERRISH and R. E. LENSKI, 1997 Evolution of high mutation rates in experimental populations of E. coli. Nature 387: 703-705.

62 SONTI, R. V., and J. R. ROTH, 1989 Role of gene duplications in the adaptation of Salmonella typhimurium to growth on limiting carbon sources. Genetics 123: 19-28. SOO, V. W., P. HANSON-MANFUL and W. M. PATRICK, 2011 Artificial gene amplification reveals an abundance of promiscuous resistance determinants in Escherichia coli. Proc Natl Acad Sci U S A 108: 1484-1489. STANTON, T. B., 2007 -like gene transfer agents-novel mechanisms of gene exchange for , , , and species. Anaerobe 13: 43-49. STRAUS, D. S., and L. D'ARI STRAUS, 1976 Large overlapping tandem genetic duplications in Salmonella typhimurium. J Mol Biol 103: 143-153. STREISINGER, G., Y. OKADA, J. EMRICH, J. NEWTON, A. TSUGITA et al., 1966 Frameshift mutations and the . This paper is dedicated to Professor Theodosius Dobzhansky on the occasion of his 66th birthday. Cold Spring Harb Symp Quant Biol 31: 77-84. SUYAMA, M., and P. BORK, 2001 Evolution of prokaryotic gene order: genome rearrangements in closely related species. Trends Genet 17: 10-13. SWORDS, W. E., B. M. CANNON and W. H. BENJAMIN, JR., 1997 Avirulence of LT2 strains of Salmonella typhimurium results from a defective rpoS gene. Infect Immun 65: 2451-2453. THIOLAS, A., C. BOLLET, B. LA SCOLA, D. RAOULT and J. M. PAGÈS, 2005 Successive emergence of aerogenes strains resistant to and colistin in a patient. Antimicrob Agents Chemother 49: 1354-1358. THOMAS, C. M., and K. M. NIELSEN, 2005 Mechanisms of, and barriers to, horizontal gene transfer between bacteria. Nat Rev Microbiol 3: 711-721. THOMSON, N. R., D. J. CLAYTON, D. WINDHORST, G. VERNIKOS, S. DAVIDSON et al., 2008 Comparative genome analysis of Salmonella Enteritidis PT4 and Salmonella Gallinarum 287/91 provides insights into evolutionary and host adaptation pathways. Genome Res 18: 1624-1637. TILLIER, E. R., and R. A. COLLINS, 2000 Genome rearrangement by replication-directed translocation. Nat Genet 26: 195-197. TLSTY, T. D., A. M. ALBERTINI and J. H. MILLER, 1984 Gene amplification in the lac region of E. coli. Cell 37: 217-224. TRINH, T. Q., and R. R. SINDEN, 1991 Preferential DNA secondary structure mutagenesis in the lagging strand of replication in E. coli. Nature 352: 544-547. TRINH, T. Q., and R. R. SINDEN, 1993 The influence of primary and secondary DNA structure in deletion and duplication between direct repeats in Escherichia coli. Genetics 134: 409-422.

63 VAARA, M., T. VAARA and M. SARVAS, 1979 Decreased binding of polymyxin by polymyxin-resistant mutants of Salmonella typhimurium. J Bacteriol 139: 664-667. VEITIA, R. A., 2002 Exploring the etiology of haploinsufficiency. Bioessays 24: 175-184. VEITIA, R. A., 2004 Gene dosage balance in cellular pathways: implications for dominance and gene duplicability. Genetics 168: 569-574. WALSH, T. R., M. A. TOLEMAN, L. POIREL and P. NORDMANN, 2005 Metallo-beta-lactamases: the quiet before the storm? Clin Microbiol Rev 18: 306-325. WELDHAGEN, G. F., 2004 Integrons and beta-lactamases--a novel perspective on resistance. Int J Antimicrob Agents 23: 556-562. WENDEL, J. F., 2000 Genome evolution in polyploids. Mol Biol 42: 225-249. WERNEGREEN, J. J., and N. A. MORAN, 1999 Evidence for genetic drift in endosymbionts (Buchnera): analyses of protein-coding genes. Mol Biol Evol 16: 83-97. WESTON-HAFER, K., and D. E. BERG, 1989 Palindromy and the location of deletion endpoints in Escherichia coli. Genetics 121: 651-658. WHORISKEY, S. K., V. H. NGHIEM, P. M. LEONG, J. M. MASSON and J. H. MILLER, 1987 Genetic rearrangements and gene amplification in Escherichia coli: DNA sequences at the junctures of amplified gene fusions. Genes Dev 1: 227-237. WIEDENBECK, J., and F. M. COHAN, 2011 Origins of bacterial diversity through horizontal genetic transfer and adaptation to new ecological niches. FEMS Microbiol Rev 35: 957-976. WONG, K. K., R. J. DELEEUW, N. S. DOSANJH, L. R. KIMM, Z. CHENG et al., 2007 A comprehensive analysis of common copy-number variations in the . Am J Hum Genet 80: 91-104. WOZNIAK, R. A., and M. K. WALDOR, 2010 Integrative and conjugative elements: mosaic mobile genetic elements enabling dynamic lateral gene flow. Nat Rev Microbiol 8: 552-563. ZAPUN, A., C. CONTRERAS-MARTEL and T. VERNET, 2008 Penicillin- binding proteins and beta-lactam resistance. FEMS Microbiol Rev 32: 361-385.

64